FreshContext A context lab · Fife , home
Drawing
02 · Benchmarks
Sheet
2 of 3
Rev
A · 2026-10-01

Working drawing 02 · Run log ← Benchmarks

Run log.

Every measurement taken, newest first. 794 runs across 19 suites and 187 configurations, from 2026-08-23 to 2026-10-27.

Runs
794
Distinct graders
32
Incomplete coverage
39

Measurements taken on an Apple M1 Max with 64 GB of unified memory.
Every number here comes from a locked dataset with a recorded grader. Where a measurement is not valid, it is marked rather than removed.

Reading it Why coverage is shown

A run that scored well on half its cases is not the same as one that scored well on all of them. Collapsing the two hides harness failures inside model results. Where coverage fell below 100 percent it is marked, and those 39 runs should be read as harness problems rather than model performance.

The grader identifier is a hash of the suite source. If it changes between two runs of the same suite, they were not scored by the same code and are not directly comparable.

All runs 794 entries · filter, sort, search
FinishedSuiteModelRuntime · profileScoreCoverageGrader · dataset · config
toolcall2 toolcall/2gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extractionn/aCoverage 120/120g 4c7867512a0a37b1 d 2ecae6a15d157146 c 1db1206325f3f3d9
chainwalk chainwalk/3gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c f308cc39aac69de4
contract contract/2gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction50.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 4eeacca43f10d263
screening screening/1gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 919410d0d5236f00
tools tools/1gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction93.3%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 8ab141cd23b0364d
toolscale toolscale/1gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 906f4ea5017ba5e4
toolcall toolcall/1gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction55.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 37f08e2a274c5629
nullschema nullschema/1gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction88.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 9a44e3886e18755c
worldview worldview/2gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction97.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 4f3d3e766cc6396d
synthesis synthesis/2gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 05a57a62cae736dc
epistemics epistemics/4gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 0c7f2ae85b8ab687
enrichment enrichment/3gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction77.8%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c c3fe2834fc3aceee
grounding grounding/1gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction84.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c b886830861b5c3f2
needle needle/1gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction59.1%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c 719e72068b3b728c
screening_native screening-native/1granite-guardian-4.1-8b-q4_0-ibm-np1-c64k-b11846 llama.cpp extraction93.3%Coverage 60/60g c25afcd28397516f d 4f4dd926349b70a0 c dde9b0957de6d134
screening screening/1granite-guardian-4.1-8b-q4_0-ibm-np1-c64k-b11846 llama.cpp extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c fb57a714152830d1
screening_native screening-native/1granite-guardian-4.1-8b-q4km-ibm-np1-c64k-b11846 llama.cpp extraction88.3%Coverage 60/60g c25afcd28397516f d 4f4dd926349b70a0 c 2a2539b427fe41ed
screening screening/1granite-guardian-4.1-8b-q4km-ibm-np1-c64k-b11846 llama.cpp extraction100.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c dcbfb0f9f4d50129
screening screening/1gemma4-12b-qat-q4_0-seat-c8k-kvq4-b10893 llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c e9e1f1dde3c4ccf2
grounding grounding/1gemma4-12b-qat-q4_0-seat-c8k-kvq4-b10893 llama.cpp extraction82.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c b02e10a69dc4f210
nullschema nullschema/1gemma4-12b-qat-q4_0-seat-c8k-kvq4-b10893 llama.cpp extraction87.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c f56c2af956a5d0b7
enrichment enrichment/3gemma4-12b-qat-q4_0-seat-c8k-kvq4-b10893 llama.cpp extraction75.9%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 78c5ded3cfeb7807
tagger_nullable tagger-nullable/1gemma4-12b-qat-q4_0-seat-c8k-kvq4-b10893 llama.cpp extraction98.3%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 232df2425a1a1903
tagger tagger/2gemma4-12b-qat-q4_0-seat-c8k-kvq4-b10893 llama.cpp extraction92.9%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c ee050e77dafe4a84
tagger_nullable tagger-nullable/1gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction97.9%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 795c1cbca42da47b
tagger tagger/2gemma4-12b-qat-q4_0-np1-c64k-b11846 llama.cpp extraction93.8%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 8036139cf9c51981
screening_native screening-native/1granite-guardian-4.1-8b-q4_0-np1-c8k-kvq4-b11846 llama.cpp extraction88.3%Coverage 60/60g c25afcd28397516f d 4f4dd926349b70a0 c b485bf201e54b22e
screening screening/1granite-guardian-4.1-8b-q4_0-np1-c8k-kvq4-b11846 llama.cpp extraction90.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 2b32a68cab8dbce8
screening_native screening-native/1granite-guardian-4.1-8b-q4_0-np1-c64k-kvq4-b11846 llama.cpp extraction88.3%Coverage 60/60g c25afcd28397516f d 4f4dd926349b70a0 c 37c021d6dba9d648
screening screening/1granite-guardian-4.1-8b-q4_0-np1-c64k-kvq4-b11846 llama.cpp extraction90.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c f46a18ca1b552709
screening_native screening-native/1granite-guardian-4.1-8b-q4_0-np1-c64k-b11846 llama.cpp extraction90.0%Coverage 60/60g c25afcd28397516f d 4f4dd926349b70a0 c 37c021d6dba9d648
screening screening/1granite-guardian-4.1-8b-q4_0-np1-c64k-b11846 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c f46a18ca1b552709
toolcall2 toolcall/2spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extractionn/aCoverage 120/120g 4c7867512a0a37b1 d 2ecae6a15d157146 c 2cd897f8ed920bdf
chainwalk chainwalk/3spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 6e9183b6e3dfb175
contract contract/2spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction25.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 5b96e006022e6218
screening screening/1spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction96.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 2f28c0b25cd855b3
tools tools/1spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction80.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c eb102ab31b1fed3f
toolscale toolscale/1spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 3cdd864ed9d75224
toolcall toolcall/1spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction0.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 45961d58b222b2b6
nullschema nullschema/1spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction48.6%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 8dc1ad6df730d117
worldview worldview/2spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction20.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 8917e280c32ae81f
synthesis synthesis/2spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction28.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c deca2750b95cfb02
epistemics epistemics/4spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction72.9%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 46daddd78b43dd51
enrichment enrichment/3spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction77.8%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c a398dac440b5228c
grounding grounding/1spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction72.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c aaa3ffccbf9aaeb4
needle needle/1spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction63.6%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c 2dc8e6d07655daca
tagger_nullable tagger-nullable/1spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction88.8%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 81db34b0e26e0e30
tagger tagger/2spark-x2.5-4b-q4km-np1-c64k-b11846 llama.cpp extraction87.5%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 1ce3949b6664155c
chainwalk chainwalk/3spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 484355bded7c5ccf
contract contract/2spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction29.2%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c a4f6da9cb513bd95
screening screening/1spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 576a7775cfad15c8
tools tools/1spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction86.7%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 2aaa6d3e20791385
toolscale toolscale/1spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c b55c738a4fe04d02
toolcall toolcall/1spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction0.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 53999f02f507d01c
nullschema nullschema/1spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction63.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 9f423724838731e4
worldview worldview/2spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c d61317fe3d8dbd88
synthesis synthesis/2spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction81.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c d06c8b2257d089b2
epistemics epistemics/4spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction95.8%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 1174e5574995ebe3
enrichment enrichment/3spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction85.2%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c e19077cf9cdc909a
grounding grounding/1spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction72.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c b09bb0071f669f4f
needle needle/1spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction87.3%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c b9316008e0cd79fa
tagger_nullable tagger-nullable/1spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction71.7%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c c3ded5d798202b34
tagger tagger/2spark-x2.6-4b-q4km-np1-c64k-b11846 llama.cpp extraction87.9%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c ce362973f5721739
chainwalk chainwalk/3lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c e7c8396d9ed03203
contract contract/2lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction75.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c d983a1eab242a2c5
screening screening/1lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction25.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c b74d2f0fd98f854b
tools tools/1lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction86.7%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 11f1bcab9b59dbbb
toolscale toolscale/1lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c dc258c69347efb48
toolcall toolcall/1lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 81b819c35899e092
nullschema nullschema/1lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction44.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c fcb6d784f212ea90
worldview worldview/2lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction55.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c ecefd0dc821be544
synthesis synthesis/2lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction75.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 3b8b41b1a21cb97b
epistemics epistemics/4lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction70.8%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c e79f2ead72c51e1f
enrichment enrichment/3lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction51.9%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c ce56e16d2c0e729c
chainwalk chainwalk/3flint-q4_1-np1-c64k-b11846 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 88345f03f0e327c4
contract contract/2flint-q4_1-np1-c64k-b11846 llama.cpp extraction60.4%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 43204cdedbb980ca
screening screening/1flint-q4_1-np1-c64k-b11846 llama.cpp extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 3178f115c8d05533
tools tools/1flint-q4_1-np1-c64k-b11846 llama.cpp extraction100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c c2432e96b7bff4bf
toolscale toolscale/1flint-q4_1-np1-c64k-b11846 llama.cpp extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c ca6e6648773e72ef
toolcall toolcall/1flint-q4_1-np1-c64k-b11846 llama.cpp extraction55.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 956656f55d16d854
nullschema nullschema/1flint-q4_1-np1-c64k-b11846 llama.cpp extraction70.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 6f004eb5725740b2
worldview worldview/2flint-q4_1-np1-c64k-b11846 llama.cpp extraction75.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 6bd294468f0eb16d
synthesis synthesis/2flint-q4_1-np1-c64k-b11846 llama.cpp extraction78.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c cfd189f0cc9137d8
epistemics epistemics/4flint-q4_1-np1-c64k-b11846 llama.cpp extraction95.8%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c ab518cd898f25ee8
enrichment enrichment/3flint-q4_1-np1-c64k-b11846 llama.cpp extraction77.8%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 6a8259b35c6b5b0d
grounding grounding/1flint-q4_1-np1-c64k-b11846 llama.cpp extraction77.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c abfb64f4eb2cecaf
needle needle/1flint-q4_1-np1-c64k-b11846 llama.cpp extraction60.0%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c 121389ab32a52e87
tagger_nullable tagger-nullable/1flint-q4_1-np1-c64k-b11846 llama.cpp extraction60.8%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c d2430ea674cbd7d9
tagger tagger/2flint-q4_1-np1-c64k-b11846 llama.cpp extraction82.1%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c abb95e6a254a1a93
grounding grounding/1lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction72.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 2ba60c7c54da5b59
needle needle/1lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction54.5%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c 65a31ff3750251f7
tagger_nullable tagger-nullable/1lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction50.0%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 36c31b70aa553f7e
tagger tagger/2lfm2.5-24b-a2b-qad-q4_0-np1-c64k-b11846 llama.cpp extraction62.1%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 1fe2084ade408d11
tagger tagger/2flint-q4_1-np4-b11846 llama.cpp extraction54.2%Coverage 10% coverageg 63e04d139a0e5ed3 d 7ef587bf33e5788a c fcbc238912075840
needle needle/1flint-q4_1-np4-b11846 llama.cpp extraction85.7%Coverage 6.4% coverageg c38dec33aa1c86c2 d 5c5199b9eee8a595 c d4b82caabaad3205
tagger_nullable tagger-nullable/1flint-q4_1-np4-b11846 llama.cpp extraction57.1%Coverage 11.7% coverageg 04e875431e26c4b5 d db3addf23a2bc4ce c 5dd5624f007db447
grounding grounding/1flint-q4_1-np4-b11846 llama.cpp extraction66.1%Coverage 44.8% coverageg 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 48d85d169f54be82
chainwalk chainwalk/3lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 1aa3b1ae71955ea1
enrichment enrichment/3lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction38.9%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c d5bc927e67bb57cc
tagger tagger/2lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction62.5%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c a483f88f4f3e2b34
screening screening/1lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction48.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 99bb653fb40dd11b
contract contract/2lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction72.9%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 66489f7c9323f5c1
tools tools/1lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction86.7%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 260296041a390e31
toolscale toolscale/1lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c a8f3c0fe0f501532
toolcall toolcall/1lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 924b8542372c5a6e
worldview worldview/2lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction27.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 268f16986415e52a
nullschema nullschema/1lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction38.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 10a03dd6e47e9d1b
synthesis synthesis/2lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction70.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 89397895ec36c956
tagger_nullable tagger-nullable/1lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction45.4%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 260eec9c0d3b6ebc
epistemics epistemics/4lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction75.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 39d2305b9089ea3b
grounding grounding/1lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction71.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 8a1861d31a5df2fd
needle needle/1lfm2.5-24b-a2b-qad-q4_0-np4-b11846 llama.cpp extraction56.4%Coverage 50% coverageg c38dec33aa1c86c2 d 5c5199b9eee8a595 c 0d0138cc0d658e0d
needle needle/1sakura-iq3xxs-np2-128k-b11533 llama.cpp extraction65.1% voidCoverage 75.5% coverageg c38dec33aa1c86c2 d 5c5199b9eee8a595 c c420a150c3bfc6d4
screening screening/1saluki-iq2mix-np4-b11552 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 7712f88cb7880b4a
synthesis synthesis/2saluki-iq2mix-np4-b11552 llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 4f33ae0ff64b3a78
toolscale toolscale/1saluki-iq2mix-np4-b11552 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c d97508b704d7aad6
worldview worldview/2saluki-iq2mix-np4-b11552 llama.cpp extraction75.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c b6e3fc11216c3462
toolcall toolcall/1saluki-iq2mix-np4-b11552 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 6dadd4e098190535
grounding grounding/1saluki-iq2mix-np4-b11552 llama.cpp extraction68.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c e0a0516acff36408
nullschema nullschema/1saluki-iq2mix-np4-b11552 llama.cpp extraction29.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 97d4f923523d65bf
needle needle/1saluki-iq2mix-np4-b11552 llama.cpp extraction68.2% voidCoverage 22/22g c38dec33aa1c86c2 d 5c5199b9eee8a595 c 2004762934080efa
tagger tagger/2saluki-iq2mix-np4-b11552 llama.cpp extraction100.0% voidCoverage 3/3g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 65e17c62027e1f61
tagger_nullable tagger-nullable/1saluki-iq2mix-np4-b11552 llama.cpp extraction100.0% voidCoverage 4/4g 04e875431e26c4b5 d db3addf23a2bc4ce c 01f66a327b74ef1c
tagger tagger/2sakura-iq3xxs-np4-b11533 llama.cpp extraction67.1%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 9d95c67412953893
tagger_nullable tagger-nullable/1sakura-iq3xxs-np4-b11533 llama.cpp extraction85.0%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c dc828964a698788d
chainwalk chainwalk/3sakura-iq3xxs-np4-b11533 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 0d5a0262425da9db
screening screening/1sakura-iq3xxs-np4-b11533 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 6c0b1e435d00872a
contract contract/2sakura-iq3xxs-np4-b11533 llama.cpp extraction75.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 702b35969843f4b8
toolscale toolscale/1sakura-iq3xxs-np4-b11533 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c e5439a083fbfffa3
tools tools/1sakura-iq3xxs-np4-b11533 llama.cpp extraction86.7%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c f78eedef33b88831
toolcall toolcall/1sakura-iq3xxs-np4-b11533 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 5c9ba80773208a76
worldview worldview/2sakura-iq3xxs-np4-b11533 llama.cpp extraction65.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c ec8cadc0716018ec
nullschema nullschema/1sakura-iq3xxs-np4-b11533 llama.cpp extraction33.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c cc113e9cf29de168
synthesis synthesis/2sakura-iq3xxs-np4-b11533 llama.cpp extraction56.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c ef20aecf8ec2da7a
enrichment enrichment/3sakura-iq3xxs-np4-b11533 llama.cpp extraction83.3%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c cc82d17507d5e18b
epistemics epistemics/4sakura-iq3xxs-np4-b11533 llama.cpp extraction100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 4f893c122bcbebd6
grounding grounding/1sakura-iq3xxs-np4-b11533 llama.cpp extraction79.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 2e91eb6cf25a79f4
needle needle/1sakura-iq3xxs-np4-b11533 llama.cpp extraction63.6% voidCoverage 50% coverageg c38dec33aa1c86c2 d 5c5199b9eee8a595 c b7246ba968262208
tagger_nullable tagger-nullable/1gemma4-12b-b11552 llama.cpp extraction95.8%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 3177d84e7059b9ff
tagger tagger/2gemma4-12b-b11552 llama.cpp extraction92.5%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 32acbf2c5241a24c
screening screening/1granite-guardian-4.1-8b-b11552-nocoh llama.cpp extraction90.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 6699fbb941461225
screening screening/1granite-guardian-4.1-8b-b10893-control-nocoh llama.cpp extraction88.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 9078bea68397a176
nullschema nullschema/1qwen3-8b-b11552 llama.cpp extraction47.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 8f248f641f1573f2
nullschema nullschema/1k2-3.7b-b11552-noreasonfmt llama.cpp extraction19.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 45d240a5f6c250ef
nullschema nullschema/1k2-3.7b-fork-control-nocoh llama.cpp extraction16.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 5eaaeb2b555f45d5
nullschema nullschema/1k2-3.7b-b11552-nocoh llama.cpp extraction15.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 45d240a5f6c250ef
tools tools/1k2-3.7b-b11552-nocoh llama.cpp extraction86.7%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c c8fce8a55fb148ae
toolcall toolcall/1k2-3.7b-b11552-nocoh llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 902aca0deb3c8313
worldview worldview/2gemma4-e2b llama.cpp extraction40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c de57b013a181e52f
worldview worldview/2gemma4-e4b llama.cpp extraction40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c d4e649bd3a7302b4
chainwalk chainwalk/3kolibri-1-q4_0 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 3adb18730a2383fc
nullschema nullschema/1kolibri-1-q4_0 llama.cpp extraction54.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 9c135c7598144ea1
worldview worldview/2kolibri-1-q4_0 llama.cpp extraction60.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 4f1168c31c8353b6
synthesis synthesis/2kolibri-1-q4_0 llama.cpp extraction63.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c cfa9291c3acf17db
screening screening/1kolibri-1-q4_0 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 0fedfa80c8db5e20
toolscale toolscale/1kolibri-1-q4_0 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 6895869428b37b03
toolcall toolcall/1kolibri-1-q4_0 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 1e106aea3ca6763c
toolscale toolscale/1qwen3.6-a3b-splash-nothink splash study-nothink77.5%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 771acc600d2fc5d4
epistemics epistemics/4qwen3.6-a3b-splash-nothink splash study-nothink100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c f6c0f7aeb3402cfd
toolcall toolcall/1qwen3.6-a3b-splash-nothink splash study-nothink56.7%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c efb08f5f59706d55
worldview worldview/2qwen3.6-a3b-splash-nothink splash study-nothink40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 8f498d3c990a6efd
synthesis synthesis/2qwen3.6-a3b-splash-nothink splash study-nothink78.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c d39bd54b99600317
grounding grounding/1qwen3.6-a3b-splash-nothink splash study-nothink70.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 6ede503e649838d4
nullschema nullschema/1qwen3.6-a3b-splash-nothink splash study-nothink34.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c bcff3aee003690c6
toolscale toolscale/1qwen3.8-27b-splash-nothink splash study-nothink100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c ca9501ceb8f24634
epistemics epistemics/4qwen3.8-27b-splash-nothink splash study-nothink100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 0af5e5369eb48698
toolcall toolcall/1qwen3.8-27b-splash-nothink splash study-nothink100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 6c12a51148cde6cf
worldview worldview/2qwen3.8-27b-splash-nothink splash study-nothink80.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 1a0902377d6c06c8
synthesis synthesis/2qwen3.8-27b-splash-nothink splash study-nothink83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 93b6fae8e6a15e8a
grounding grounding/1qwen3.8-27b-splash-nothink splash study-nothink83.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c a9d93d67a31bbab3
nullschema nullschema/1qwen3.8-27b-splash-nothink splash study-nothink51.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 5ae32fbfa294e0bf
toolscale toolscale/1qwen3.8-27b-splash splash study-nothink95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c ca9501ceb8f24634
epistemics epistemics/4qwen3.8-27b-splash splash study-nothink100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 0af5e5369eb48698
toolcall toolcall/1qwen3.8-27b-splash splash study-nothink100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 6c12a51148cde6cf
worldview worldview/2qwen3.8-27b-splash splash study-nothink97.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 1a0902377d6c06c8
synthesis synthesis/2qwen3.8-27b-splash splash study-nothink88.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 93b6fae8e6a15e8a
grounding grounding/1qwen3.8-27b-splash splash study-nothink84.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c a9d93d67a31bbab3
nullschema nullschema/1qwen3.8-27b-splash splash study-nothink75.0%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 5ae32fbfa294e0bf
tagger tagger/2ornith-35b-tagger llama.cpp extraction85.8%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 02e7b350ee762892
needle needle/1edge0-35b mlx_lm extraction62.7%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c cc08042253cc6e13
grounding grounding/1bonsai2-27b-ptq1 llama.cpp extraction66.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 0407c0f9d876e4a3
epistemics epistemics/4bonsai2-27b-ptq1 llama.cpp study-nothink100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 9255c41b05f7da62
screening screening/1bonsai2-27b-ptq1 llama.cpp extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 2f2913b42c03fae2
worldview worldview/2bonsai2-27b-ptq1 llama.cpp extraction82.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 40c68eea7655c7e1
screening screening/1qwen38-27b-q3kxl-chad mlx_lm extraction84.8% voidCoverage 55% coverageg 0661b49a10be2d2a d 64c97ca2269c29af c e4109198ec0d6a3f
worldview worldview/2qwen38-27b-q3kxl-chad mlx_lm extraction92.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 7f85527901732b16
synthesis synthesis/2qwen38-27b-q3kxl-chad mlx_lm extraction85.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c a1f473e0d63ec1b3
worldview worldview/2bonsai2-27b-ptq1 llama.cpp extraction100.0% voidCoverage 9/9g 9acfba1b704feb28 d e97835de9e6fc227 c 40c68eea7655c7e1
toolscale toolscale/1bonsai2-27b-ptq1 llama.cpp agentic-fast97.5%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 92433e8a2c1d7275
toolcall toolcall/1bonsai2-27b-ptq1 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 3e70e301928f2a3d
synthesis synthesis/2bonsai2-27b-ptq1 llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 522c48da0b9474a3
nullschema nullschema/1bonsai2-27b-ptq1 llama.cpp extraction31.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 2641ce03c27a545e
grounding grounding/1bonsai2-27b-pq2 llama.cpp extraction66.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 1685634443e1889f
epistemics epistemics/4bonsai2-27b-pq2 llama.cpp study-nothink100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 3e65308eee2e3c69
screening screening/1bonsai2-27b-pq2 llama.cpp extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 75d8d5feb45c7b60
worldview worldview/2bonsai2-27b-pq2 llama.cpp extraction82.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 61b04f512aeac639
toolscale toolscale/1bonsai2-27b-pq2 llama.cpp agentic-fast97.5%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c c76ee9534d06b617
toolcall toolcall/1bonsai2-27b-pq2 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c fa07c48a3877570f
synthesis synthesis/2bonsai2-27b-pq2 llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 25c89df1304600f0
nullschema nullschema/1bonsai2-27b-pq2 llama.cpp extraction31.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 537688bcd022a4e2
screening screening/1qwen3.6-35b-swiftlm swiftlm extraction88.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 7d05a9cf37c57f7c
worldview worldview/2ornith-35b-a3b-q40-recheck llama.cpp extraction85.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 2124361885388e32
screening screening/1edge0-35b-recheck mlx_lm extraction88.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c f685f626196f2883
screening screening/1ternary-bonsai-27b-recheck mlx_lm extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c a5afffb46b0e50c4
screening screening/1qwen3.8-9b-distill-recheck llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 4b6dbb06475bbbad
screening screening/1gemma4-12b-recheck llama.cpp extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 321d8d33ba14ed0c
screening screening/1btl-4-q4km-recheck llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 86e42deccfe15a63
tagger tagger/2minicpm5-2b-q8-nojinja llama.cpp extraction76.9%Coverage 99.2% coverageg 63e04d139a0e5ed3 d 7ef587bf33e5788a c 0153d351e0ec30ee
tagger tagger/2minicpm5-2b-q8-np4-today llama.cpp extraction76.9%Coverage 99.2% coverageg 63e04d139a0e5ed3 d 7ef587bf33e5788a c 0d285fb9372dab00
worldview worldview/2minicpm5-2b-q8-np1 llama.cpp extraction57.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c ff823de24a74e591
synthesis synthesis/2minicpm5-2b-q8-np1 llama.cpp extraction78.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c e59bb419db23bea7
grounding grounding/1minicpm5-2b-q8-np1 llama.cpp extraction56.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 5ac3f164bce7795b
enrichment enrichment/3minicpm5-2b-q8-np1 llama.cpp extraction57.4%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 1fb5d909364039d2
tagger_nullable tagger-nullable/1minicpm5-2b-q8-np1 llama.cpp extraction52.9%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 070be56608b35405
screening screening/1minicpm5-2b-q8-np1 llama.cpp extraction83.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 12f2ce70461eb2ba
epistemics epistemics/4minicpm5-2b-q8-np1 llama.cpp study-nothink79.2%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c d8638e6e2d8d038b
tagger tagger/2minicpm5-2b-q8-np1 llama.cpp extraction76.7%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c c1fd37845b5547c8
worldview worldview/2minicpm5-2b-exl3 exl3 extraction37.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c f0bd4498f4e1a804
synthesis synthesis/2minicpm5-2b-exl3 exl3 extraction75.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c d51ee21c635d7430
needle needle/1minicpm5-2b-exl3 exl3 extraction85.5%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c 2b303f0a51347e71
grounding grounding/1minicpm5-2b-exl3 exl3 extraction60.0%Coverage 80% coverageg 30b933dcc6a7e2d4 d dcb396faf8b83b0a c deed1276a964eb4d
enrichment enrichment/3minicpm5-2b-exl3 exl3 extraction59.3%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 50da51be20c4e83c
tagger_nullable tagger-nullable/1minicpm5-2b-exl3 exl3 extraction46.2%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c dfef0a54e96403a1
tagger tagger/2minicpm5-2b-exl3 exl3 extraction76.2%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 58ec2a66d452536c
screening screening/1minicpm5-2b-exl3 exl3 extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 1564c01170861ff4
epistemics epistemics/4minicpm5-2b-exl3 exl3 study-nothink81.2%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 18bb944eee207790
tagger_nullable tagger-nullable/1ternary-bonsai-8b llama.cpp extraction61.2%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 1a6187eaf0f19f79
tagger tagger/2ternary-bonsai-8b llama.cpp extraction43.8%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 1b7169aea80ec684
tagger tagger/2ternary-bonsai-8b llama.cpp extraction36.4% voidCoverage 11/11g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 1b7169aea80ec684
enrichment enrichment/3ternary-bonsai-8b llama.cpp extraction70.4%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 2449c0a03ec229c7
contract contract/2ternary-bonsai-8b llama.cpp extraction22.9%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c b5bcdf7e22a61983
epistemics epistemics/4ternary-bonsai-8b llama.cpp study-nothink72.9%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 22502c57cf79b73d
grounding grounding/1ternary-bonsai-8b llama.cpp extraction51.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c a964482c0a6fefb9
screening screening/1ternary-bonsai-8b llama.cpp extraction86.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 40d9d3c65e78e666
synthesis synthesis/2ternary-bonsai-8b llama.cpp extraction60.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 00939f8fd32a28ff
worldview worldview/2ternary-bonsai-8b llama.cpp extraction37.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 0b08940f0421efe8
nullschema nullschema/1ternary-bonsai-8b llama.cpp extraction11.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 409ab3394165e872
tools tools/1ternary-bonsai-8b llama.cpp agentic-fast93.3%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c d3347a668da25f1d
toolscale toolscale/1ternary-bonsai-8b llama.cpp agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 4a4dea00c5224630
toolcall toolcall/1ternary-bonsai-8b llama.cpp extraction0.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 2ffaff38658dd85a
synthesis synthesis/2minicpm5-2b-exl3 exl3 extraction100.0% voidCoverage 21.7% coverageg 4156dd3c13361e8b d 267d1954e225c5b6 c c0f3ff2bda3a17e8
worldview worldview/2minicpm5-2b-exl3 exl3 extraction80.0% voidCoverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c d413710f2e637004
tagger_nullable tagger-nullable/1edge0-35b edge0 extraction91.7%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 7b18c32f56523945
tagger tagger/2edge0-35b edge0 extraction88.3%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 7149c1e662a034c9
tagger tagger/2edge0-35b edge0 extraction85.5% voidCoverage 83/83g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 7149c1e662a034c9
synthesis synthesis/2edge0-35b edge0 extraction81.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 0955df95170a937b
screening screening/1edge0-35b edge0 extraction88.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 60a41f7096ed2f0d
enrichment enrichment/3edge0-35b edge0 extraction77.8%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c a0c2477ba575e774
epistemics epistemics/4edge0-35b edge0 study-nothink97.9%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 63bebcfa1d0c9c54
worldview worldview/2edge0-35b edge0 extraction52.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c c4a49113da942eeb
grounding grounding/1edge0-35b edge0 extraction89.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c eb1910ec33534645
tagger_nullable tagger-nullable/1minicpm5-2b llama.cpp extraction22.9%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 8d6eef850a9bedcc
tagger tagger/2minicpm5-2b llama.cpp extraction51.7%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 8bc8889aa4bec4ee
enrichment enrichment/3minicpm5-2b llama.cpp extraction59.3%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c a602b52d4aa5d258
contract contract/2minicpm5-2b llama.cpp extraction70.8%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 09ef74ce83b54a17
epistemics epistemics/4minicpm5-2b llama.cpp study-nothink77.1%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 8bf3c6f0ca1fa01a
grounding grounding/1minicpm5-2b llama.cpp extraction62.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 8f32a36bd07ac12d
screening screening/1minicpm5-2b llama.cpp extraction85.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 859326c246bacb4d
synthesis synthesis/2minicpm5-2b llama.cpp extraction60.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 602e85a1c355f821
worldview worldview/2minicpm5-2b llama.cpp extraction37.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c d8fc1797ed992595
nullschema nullschema/1minicpm5-2b llama.cpp extraction31.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c a792a14d7d990d1b
tools tools/1minicpm5-2b llama.cpp agentic-fast93.3%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 29d5dc1655f1e02d
toolscale toolscale/1minicpm5-2b llama.cpp agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 2400bd439331eb46
toolcall toolcall/1minicpm5-2b llama.cpp extraction73.3%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c d5030ed4472a8b63
tagger_nullable tagger-nullable/1btl-4-q4km llama.cpp extraction71.3%Coverage 68.3% coverageg 04e875431e26c4b5 d db3addf23a2bc4ce c 53f41ca547c1a228
tagger_nullable tagger-nullable/1btl-4-q4km llama.cpp extraction71.3% voidCoverage 164/164g 04e875431e26c4b5 d db3addf23a2bc4ce c 53f41ca547c1a228
tagger tagger/2btl-4-q4km llama.cpp extraction87.1%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 10dcbcd1b3bb7b91
enrichment enrichment/3btl-4-q4km llama.cpp extraction81.5%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 3b3b53961682ee7f
contract contract/2btl-4-q4km llama.cpp extraction70.8%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c c5f63e89ad62fa86
chainwalk chainwalk/3btl-4-q4km llama.cpp agentic-fast0.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 9b7c71d675210c91
epistemics epistemics/4btl-4-q4km llama.cpp study-nothink100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 0689f579cb5fa875
grounding grounding/1btl-4-q4km llama.cpp extraction76.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c b6421e71dc1b07d5
screening screening/1btl-4-q4km llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c df6b9c70d3ecb99c
synthesis synthesis/2btl-4-q4km llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 34ecdb4bf4ded955
worldview worldview/2btl-4-q4km llama.cpp extraction55.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 7cf929c72ed89ccb
nullschema nullschema/1btl-4-q4km llama.cpp extraction44.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 43c9232ce793d709
tools tools/1btl-4-q4km llama.cpp agentic-fast100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c a04248bbed30bba3
toolscale toolscale/1btl-4-q4km llama.cpp agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 83afd32cbd70262a
toolcall toolcall/1btl-4-q4km llama.cpp extraction93.3%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 0adbf9c03a0169c7
tagger_nullable tagger-nullable/1btl-4 llama.cpp extraction91.7%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c f9c64867e9dfb91b
tagger tagger/2btl-4 llama.cpp extraction85.0%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 81d3e449bf19eaea
enrichment enrichment/3btl-4 llama.cpp extraction87.0%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 0b71856ba217c91c
contract contract/2btl-4 llama.cpp extraction64.6%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c bf1f8546a0e1b97c
chainwalk chainwalk/3btl-4 llama.cpp agentic-fast0.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 8dbbe23a8427e3bc
epistemics epistemics/4btl-4 llama.cpp study-nothink100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 438aa2f62c6153af
grounding grounding/1btl-4 llama.cpp extraction76.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 7d9f8293fd874d62
screening screening/1btl-4 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 8197a1913a41df95
synthesis synthesis/2btl-4 llama.cpp extraction80.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 11c6d6217d23f0b7
worldview worldview/2btl-4 llama.cpp extraction62.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 0104404569f7c535
nullschema nullschema/1btl-4 llama.cpp extraction36.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 9b2dcb073c569fce
tools tools/1btl-4 llama.cpp agentic-fast100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c e43140ad3e77d008
toolscale toolscale/1btl-4 llama.cpp agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 962089edc90681de
toolcall toolcall/1btl-4 llama.cpp extraction76.7%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 7d3d7741437436e1
tagger_nullable tagger-nullable/1bonsai-27b-1bit llama.cpp extraction90.8%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c f3c4a61093079a3e
tagger tagger/2bonsai-27b-1bit llama.cpp extraction85.4%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c d75dadae6c07b420
tagger_nullable tagger-nullable/1bonsai-27b-mlx2-cortext mlx_lm extraction88.8%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 67cfcfc5c61806c7
tagger tagger/2bonsai-27b-mlx2-cortext mlx_lm extraction87.5%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c a7bdc6604033d727
tagger_nullable tagger-nullable/1bonsai-8b-1bit llama.cpp extraction41.7%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c d5d5bc2627ca6b54
enrichment enrichment/3bonsai-27b-1bit llama.cpp extraction79.6%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 612035dbfdcf9143
tagger tagger/2bonsai-8b-1bit llama.cpp extraction28.3%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c e931e287089b3229
contract contract/2bonsai-27b-1bit llama.cpp extraction25.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c f3e9344f382b675d
chainwalk chainwalk/3bonsai-27b-1bit llama.cpp agentic-fast0.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 315259486d2b1e0e
enrichment enrichment/3bonsai-27b-mlx2-cortext mlx_lm extraction72.2%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c c6981047f9ae77d3
epistemics epistemics/4bonsai-27b-1bit llama.cpp study-nothink87.5%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 2f1d2e8b3a422ec6
contract contract/2bonsai-27b-mlx2-cortext mlx_lm extraction75.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 1db1392d6a3fe16b
grounding grounding/1bonsai-27b-1bit llama.cpp extraction67.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 0f1cdb00af0adb60
tagger_nullable tagger-nullable/1bonsai-27b-mlx2-m4pro unknown extraction89.2%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 29c36e514c29b1c2
screening screening/1bonsai-27b-1bit llama.cpp extraction96.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 06fe499419357acb
enrichment enrichment/3bonsai-8b-1bit llama.cpp extraction61.1%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c e43f51321b57c883
synthesis synthesis/2bonsai-27b-1bit llama.cpp extraction35.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c b6b3f7a5739fbaf7
worldview worldview/2bonsai-27b-1bit llama.cpp extraction40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c c66967b2cb84959c
nullschema nullschema/1bonsai-27b-1bit llama.cpp extraction18.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c b6bf04bcfeb45185
contract contract/2bonsai-8b-1bit llama.cpp extraction20.8%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c a575849c8bd49029
chainwalk chainwalk/3bonsai-8b-1bit llama.cpp agentic-fast0.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 6d222e16c742210f
tools tools/1bonsai-27b-1bit llama.cpp agentic-fast100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c a9cb56f68c317ce8
toolscale toolscale/1bonsai-27b-1bit llama.cpp agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 8ee047a8f4681484
epistemics epistemics/4bonsai-8b-1bit llama.cpp study-nothink79.2%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c f1ce82d657bdf4e2
grounding grounding/1bonsai-8b-1bit llama.cpp extraction55.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 1ac5570cbae78ecd
screening screening/1bonsai-8b-1bit llama.cpp extraction86.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 7b113b84e3e9ea89
synthesis synthesis/2bonsai-8b-1bit llama.cpp extraction65.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 49c73224f7b988ec
worldview worldview/2bonsai-8b-1bit llama.cpp extraction40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 1f1610844ef40acb
nullschema nullschema/1bonsai-8b-1bit llama.cpp extraction19.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c de1a0f97f9063521
tools tools/1bonsai-8b-1bit llama.cpp agentic-fast93.3%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 79e4f8448168d640
toolscale toolscale/1bonsai-8b-1bit llama.cpp agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 6fc2b443c7d76dde
tagger tagger/2bonsai-27b-mlx2-m4pro unknown extraction87.9%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 821ca1ce68849b25
toolcall toolcall/1bonsai-27b-1bit llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 0f6815eef83b64d0
toolcall toolcall/1bonsai-8b-1bit llama.cpp extraction0.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 516c9c0e0de06220
epistemics epistemics/4bonsai-27b-mlx2-cortext mlx_lm study-nothink89.6%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c f8818645d3ba1522
grounding grounding/1bonsai-27b-mlx2-cortext mlx_lm extraction71.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 4f595a317a905809
screening screening/1bonsai-27b-mlx2-cortext mlx_lm extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 64e14fdadbcea0e6
enrichment enrichment/3bonsai-27b-mlx2-m4pro unknown extraction72.2%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c e41b038c77c413b8
synthesis synthesis/2bonsai-27b-mlx2-cortext mlx_lm extraction66.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c a3e244ab46faf1d3
contract contract/2bonsai-27b-mlx2-m4pro unknown extraction75.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 5836be70b11516a5
worldview worldview/2bonsai-27b-mlx2-cortext mlx_lm extraction42.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 88e5e9056622def0
tools tools/1bonsai-27b-mlx2-cortext mlx_lm agentic-fast100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 8c56b2e4e619015f
toolscale toolscale/1bonsai-27b-mlx2-cortext mlx_lm agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c bf7830ad8b1e5adb
toolcall toolcall/1bonsai-27b-mlx2-cortext mlx_lm extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 072e0651b5a1976c
chainwalk chainwalk/3bonsai-27b-mlx2-m4pro mlx_lm agentic-fast0.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 16fb378c7bf6e96a
tagger_nullable tagger-nullable/1ternary-bonsai-27b llama.cpp extraction90.0%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 29960d736fce3ea5
tagger tagger/2ternary-bonsai-27b llama.cpp extraction85.8%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 773ed13c720ca034
enrichment enrichment/3ternary-bonsai-27b llama.cpp extraction72.2%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c 075725332e318d6b
epistemics epistemics/4bonsai-27b-mlx2-m4pro mlx_lm study-nothink89.6%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 3c03af0d96b59518
contract contract/2ternary-bonsai-27b llama.cpp extraction75.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c e5777a35ca1e5f10
grounding grounding/1bonsai-27b-mlx2-m4pro mlx_lm extraction71.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 86fb3d4c7b477c6c
chainwalk chainwalk/3ternary-bonsai-27b llama.cpp agentic-fast0.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 078f17fd21ba7d57
screening screening/1bonsai-27b-mlx2-m4pro mlx_lm extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 4082207a1256c4b6
synthesis synthesis/2bonsai-27b-mlx2-m4pro mlx_lm extraction66.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 14f4738d0f91fc84
worldview worldview/2bonsai-27b-mlx2-m4pro mlx_lm extraction42.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c a4f61e2b7e694288
tools tools/1bonsai-27b-mlx2-m4pro mlx_lm agentic-fast100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 98ea18969989ee4e
toolscale toolscale/1bonsai-27b-mlx2-m4pro mlx_lm agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 52e205dd25e5567b
epistemics epistemics/4ternary-bonsai-27b llama.cpp study-nothink85.4%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 321e00a36a5faaeb
grounding grounding/1ternary-bonsai-27b llama.cpp extraction68.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 3b0afe72fb0c0121
toolcall toolcall/1bonsai-27b-mlx2-m4pro mlx_lm extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 8e73d545a7ffd2aa
screening screening/1ternary-bonsai-27b llama.cpp extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c c5097500538c3719
synthesis synthesis/2ternary-bonsai-27b llama.cpp extraction66.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 6a368a22cfb7fb02
worldview worldview/2ternary-bonsai-27b llama.cpp extraction42.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 3a4970c1d0eed32d
nullschema nullschema/1ternary-bonsai-27b llama.cpp extraction27.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 004a863515b30c86
tools tools/1ternary-bonsai-27b llama.cpp agentic-fast100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 7f71a1d06f005b6d
toolscale toolscale/1ternary-bonsai-27b llama.cpp agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 13185468aa0f107f
toolcall toolcall/1ternary-bonsai-27b llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 7dece85f852d6798
tagger_nullable tagger-nullable/1qwen3.8-9b-distill llama.cpp extraction90.0%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c d7923863e8b56e76
tagger tagger/2qwen3.8-9b-distill llama.cpp extraction83.3%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c d2642a91c9a03367
enrichment enrichment/3qwen3.8-9b-distill llama.cpp extraction79.6%Coverage 54/54g b51222c2ca38cfdf d 75361864baa727ab c d537736901f01a81
contract contract/2qwen3.8-9b-distill llama.cpp extraction25.0%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 5a052624057c52ff
chainwalk chainwalk/3qwen3.8-9b-distill llama.cpp agentic-fast0.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 43d9a11df6da4927
epistemics epistemics/4qwen3.8-9b-distill llama.cpp study-nothink89.6%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c a0e9adc5c6542d53
grounding grounding/1qwen3.8-9b-distill llama.cpp extraction67.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 696fd8b30eed53f0
screening screening/1qwen3.8-9b-distill llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 5ac25f36beb78f8c
synthesis synthesis/2qwen3.8-9b-distill llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 8481e2a6e1f3e60c
worldview worldview/2qwen3.8-9b-distill llama.cpp extraction40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 6a67c959e17c98f5
nullschema nullschema/1qwen3.8-9b-distill llama.cpp extraction18.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 56c7dc8f982d59d3
tools tools/1qwen3.8-9b-distill llama.cpp agentic-fast100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 2411d23420a9fcb7
toolscale toolscale/1qwen3.8-9b-distill llama.cpp agentic-fast100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 13a2f2f87c9bcdc8
toolcall toolcall/1qwen3.8-9b-distill llama.cpp extraction0.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 2cd02d2d291d660a
epistemics epistemics/4qwen3.5-4b-base llama.cpp study-nothink85.4%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 211ae40d03020e0e
tagger_nullable tagger-nullable/1qwen3.5-4b-base llama.cpp extraction93.8%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 50a352dbe249d795
worldview worldview/2qwen3.5-4b-base llama.cpp extraction40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 314d1643f96e91a9
grounding grounding/1qwen3.5-4b-base llama.cpp extraction74.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 0f7dbcd0b1d1afd4
nullschema nullschema/1qwen3.5-4b-base llama.cpp extraction33.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 8ed6c2d36dcf6e58
toolscale toolscale/1qwen3.5-4b-base llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 634f8ac324d31747
toolcall toolcall/1qwen3.5-4b-base llama.cpp extraction36.7%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c abdddca6a1cde4d9
epistemics epistemics/4neohorse-1-4b llama.cpp study-nothink75.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c a4f7cdc18f2e769e
tagger_nullable tagger-nullable/1neohorse-1-4b llama.cpp extraction86.2%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 78196673fb9772d8
worldview worldview/2neohorse-1-4b llama.cpp extraction40.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 0ca5e9b594862393
grounding grounding/1neohorse-1-4b llama.cpp extraction63.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c b382340e2e3ad6fa
nullschema nullschema/1neohorse-1-4b llama.cpp extraction27.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c e55e57d062c38080
toolscale toolscale/1neohorse-1-4b llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c b90d64d3c6c193a5
toolcall toolcall/1neohorse-1-4b llama.cpp extraction56.7%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c e018b8e820715109
needle needle/1ornith-9b-kv-q40 llama.cpp extraction55.5%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c 09bf0b1e19047eb2
needle needle/1ornith-9b-kv-f16 llama.cpp extraction58.2%Coverage 110/110g c38dec33aa1c86c2 d 5c5199b9eee8a595 c 72650e0b4ba22c6c
synthesis synthesis/2gemma4-12b-b10893 llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 41030e0a832a129e
epistemics epistemics/4ornith-1.5-9b llama.cpp extraction100.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 54cc298fd20d22af
contract contract/2ornith-1.5-9b llama.cpp extraction68.8%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 1bf634506e78523a
synthesis synthesis/2ornith-1.5-9b llama.cpp extraction80.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 3fba7b7b697d2bc6
nullschema nullschema/1glm-4.7-flash-q40 llama.cpp extraction44.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c a206ab6679bc0fed
epistemics epistemics/4glm-4.7-flash-q40 llama.cpp extraction75.0%Coverage 48/48g acf84972a0ad5f38 d 4bce453fd376553c c 0550121d973797df
worldview worldview/2glm-4.7-flash-q40 llama.cpp extraction37.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 9d24f8d5c84a9e2e
synthesis synthesis/2glm-4.7-flash-q40 llama.cpp extraction60.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c b33ed27337b5ef7e
toolscale toolscale/1glm-4.7-flash-q40 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 5383698d8f3d7c82
toolcall toolcall/1glm-4.7-flash-q40 llama.cpp extraction38.3%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c f492f0f6a7e78f98
toolcall toolcall/1ornith-9b-kvq4 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 970506005273665a
grounding grounding/1ornith-9b-kvq4 llama.cpp extraction80.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 4c61ed44aed10e7a
nullschema nullschema/1ornith-9b-kvq4 llama.cpp extraction56.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 32a57cca965b6777
worldview worldview/2ornith-35b-a3b-q40 llama.cpp extraction85.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 36403d17c401159a
grounding grounding/1ornith-35b-a3b-q40 llama.cpp extraction76.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 95c69b1144e96db8
nullschema nullschema/1ornith-35b-a3b-q40 llama.cpp extraction36.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 4324ac01a6993c44
contract contract/2ornith-35b-a3b-q40 llama.cpp extraction66.7%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c 16c1587da02e495a
toolscale toolscale/1ornith-35b-a3b-q40 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 9ed2b4b41dc68126
toolcall toolcall/1ornith-35b-a3b-q40 llama.cpp extraction80.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 9391215fff65e4b5
worldview worldview/2ornith-1.5-9b llama.cpp extraction62.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c ce57d773fcc0cbcd
grounding grounding/1ornith-1.5-9b llama.cpp extraction81.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 4a6ad96362484d18
nullschema nullschema/1ornith-1.5-9b llama.cpp extraction52.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 0dd3a4ce0c1655b8
toolscale toolscale/1ornith-1.5-9b llama.cpp extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 8105ecbc4488d189
toolcall toolcall/1ornith-1.5-9b llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c c2191a274d61789c
tagger_nullable tagger-nullable/1gpt-oss-20b-study-default llama.cpp study-default93.9% voidCoverage 13.8% coverageg 04e875431e26c4b5 d db3addf23a2bc4ce c 7649f91ca5eb87a4
nullschema nullschema/1gpt-oss-20b llama.cpp extraction72.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 262a6f5122bb9bbd
tagger tagger/2gpt-oss-20b llama.cpp extraction86.2%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c b9bd7f33427388fd
tagger_nullable tagger-nullable/1gpt-oss-20b llama.cpp extraction85.7%Coverage 99.2% coverageg 04e875431e26c4b5 d db3addf23a2bc4ce c acb0e0e0f1883336
nullschema nullschema/1qwen35-9b-mlx-mxfp4 rapid-mlx extraction34.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c da30e30743dadebe
nullschema nullschema/1qwen35-9b-mlx-affine-g32 rapid-mlx extraction45.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 4cb05505e1d4b652
nullschema nullschema/1qwen35-9b-mlx-affine-g64 rapid-mlx extraction59.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 987f7b78b76a6a60
nullschema nullschema/1qwen35-9b-gguf-q40 llama.cpp extraction6.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c e5fa21270174a387
nullschema nullschema/1qwen35-9b-gguf-q4km llama.cpp extraction23.6%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 71580a34709d8143
synthesis synthesis/2qwen3.8-27b-q4_0 llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 7bca9c0f60cfafa3
nullschema nullschema/1qwen3.8-27b-q4_0 llama.cpp extraction69.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 0464331c99998f53
worldview worldview/2qwen3.8-27b-q4_0 llama.cpp extraction80.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 4fb569ecc91140c9
toolscale toolscale/1qwen3.8-27b-q4_0 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 7dedd451e841393d
toolcall toolcall/1qwen3.8-27b-q4_0 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c ce7aa224f6d8c444
nullschema nullschema/1k2-36b-Q4_1 llama.cpp extraction50.0%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c a1e412de2582e21b
nullschema nullschema/1k2-36b-Q5_1 llama.cpp extraction58.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 6e470ceac8f4401c
nullschema nullschema/1k2-36b-Q5_0 llama.cpp extraction56.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 53f943deaafde391
synthesis synthesis/2k2-36b-q4_0 llama.cpp extraction76.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 4bc7f1a68df72179
nullschema nullschema/1k2-36b-q4_0 llama.cpp extraction65.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c c71d223f0ac921ec
worldview worldview/2k2-36b-q4_0 llama.cpp extraction57.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c d2167f66564312b7
toolscale toolscale/1k2-36b-q4_0 llama.cpp extraction90.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 5467852ddcdb11c6
toolcall toolcall/1k2-36b-q4_0 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 646dbb2eca06be4d
toolcall toolcall/1reap-60b-mlx-4bit-uniform rapid-mlx extraction-t0728.3%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c d696a3a98ce65e4f
toolcall toolcall/1reap-60b-4bit-repro rapid-mlx extraction25.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 9fc8dc4dd468ed18
synthesis synthesis/2reap-60b-mlx-4bit-uniform rapid-mlx extraction85.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 5f382a39bf24731a
nullschema nullschema/1reap-60b-mlx-4bit-uniform rapid-mlx extraction84.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 52349074a17dc8f5
worldview worldview/2reap-60b-mlx-4bit-uniform rapid-mlx extraction47.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 76804a68d58c5eb6
toolscale toolscale/1reap-60b-mlx-4bit-uniform rapid-mlx extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c a58faee14a0f3ed6
toolcall toolcall/1reap-60b-mlx-4bit-uniform rapid-mlx extraction25.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 9fc8dc4dd468ed18
synthesis synthesis/2diffusiongemma-26b-mlx mlx-vlm extraction80.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 6220b53efc9f35b4
worldview worldview/2diffusiongemma-26b-mlx mlx-vlm extraction77.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c b0b19b278db1058f
toolscale toolscale/1diffusiongemma-26b-mlx mlx-vlm extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 40a7e393e161d98f
toolcall toolcall/1diffusiongemma-26b-mlx mlx-vlm extraction5.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 49640f29f7104a95
synthesis synthesis/2reap-60b-gguf-q4km llama.cpp extraction81.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 5d5c05e03d9c61e2
VOID-degenerate-nullschema nullschema/1reap-60b-gguf-q4km llama.cpp extraction100.0% voidCoverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c b923bc8484d5f158
worldview worldview/2reap-60b-gguf-q4km llama.cpp extraction80.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 361b96fbc9be289c
toolscale toolscale/1reap-60b-gguf-q4km llama.cpp extraction87.5%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 437c75c01b4063cd
toolcall toolcall/1reap-60b-gguf-q4km llama.cpp extraction41.7%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 4faf8950bd5037a5
synthesis synthesis/2reap-60b-20260910 rapid-mlx extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 078d5fad4e6d0a0f
nullschema nullschema/1reap-60b-20260910 rapid-mlx extraction47.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 067da0ca87821991
worldview worldview/2reap-60b-20260910 rapid-mlx extraction47.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c a9d79c1de500e4b4
toolscale toolscale/1reap-60b-20260910 rapid-mlx extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c ce60a8f4e40a948a
toolcall toolcall/1reap-60b-20260910 rapid-mlx extraction78.3%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c d3a58130e11b1504
synthesis synthesis/2spark-x25-4b llama.cpp extraction28.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c ffe0798e394ad2c2
nullschema nullschema/1spark-x25-4b llama.cpp extraction48.6%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 080b2b956a3c69aa
worldview worldview/2spark-x25-4b llama.cpp extraction20.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 7eb3a5699dc38e49
toolscale toolscale/1spark-x25-4b llama.cpp extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 266788017096e178
toolcall toolcall/1spark-x25-4b llama.cpp extraction0.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c bd8238a4b2ea134b
worldview worldview/2qwen38-rebase-b10893 llama.cpp extraction77.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 90bd067cd59ed0c2
nullschema nullschema/1qwen38-rebase-b10893 llama.cpp extraction76.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 317df1bd5861fa04
worldview worldview/2k2-36b-q8-precision llama.cpp extraction55.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c c95b13073128b76a
nullschema nullschema/1k2-36b-q8-precision llama.cpp extraction70.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 3a41a0a4d5f18851
worldview worldview/2k2-7b-bf16-precision llama.cpp extraction30.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 1705a034b7c84210
nullschema nullschema/1k2-7b-bf16-precision llama.cpp extraction37.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c d931535c0ebf24a6
worldview worldview/2k2-7b-q8-precision llama.cpp extraction30.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c e20e4b1b964fa50d
nullschema nullschema/1k2-7b-q8-precision llama.cpp extraction36.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c ed643e52a40b0e33
worldview worldview/2k2-7b-q4-precision llama.cpp extraction37.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 3bbb4039e7a64a8b
nullschema nullschema/1k2-7b-q4-precision llama.cpp extraction37.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 577908cfcedc7d78
synthesis synthesis/2k2-4b-bf16-precision llama.cpp extraction56.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 0faa5511a9576d51
nullschema nullschema/1k2-4b-bf16-precision llama.cpp extraction43.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 776fc6d33aa8cc34
worldview worldview/2k2-4b-bf16-precision llama.cpp extraction50.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 770c2ce3d2f16e96
toolscale toolscale/1k2-4b-bf16-precision llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c e21f6ac8887f7f5d
toolcall toolcall/1k2-4b-bf16-precision llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 2c2e5971d9f6382f
synthesis synthesis/2k2-3.7b-q8-precision llama.cpp extraction56.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 3529bab4b75f7502
nullschema nullschema/1k2-3.7b-q8-precision llama.cpp extraction43.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 807203bd18193834
worldview worldview/2k2-3.7b-q8-precision llama.cpp extraction50.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c f654ce7b2d1ac877
toolscale toolscale/1k2-3.7b-q8-precision llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 1eedcfdecea95e86
toolcall toolcall/1k2-3.7b-q8-precision llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c ddb3ca71c12850c5
synthesis synthesis/2k2-3.7b-q4-precision llama.cpp extraction50.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 49e024d8aab08569
nullschema nullschema/1k2-3.7b-q4-precision llama.cpp extraction38.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c b4a841d2cc241d1f
worldview worldview/2k2-3.7b-q4-precision llama.cpp extraction52.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 324471b050d0453b
toolscale toolscale/1k2-3.7b-q4-precision llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 8bb7dfa4a8ce435f
toolcall toolcall/1k2-3.7b-q4-precision llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 9c699af63ecd43af
synthesis synthesis/2gemma4-12b-midrange llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 480148dc413153dc
nullschema nullschema/1gemma4-12b-midrange llama.cpp extraction87.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c c10333ca24baf0ea
worldview worldview/2gemma4-12b-midrange llama.cpp extraction97.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 4bbaa37911174c88
toolscale toolscale/1gemma4-12b-midrange llama.cpp extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c bf8723aa88c0d4b7
toolcall toolcall/1gemma4-12b-midrange llama.cpp extraction55.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 57bdac944281e060
synthesis synthesis/2lfm25-8b-a1b-midrange llama.cpp extraction53.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 089fdc592c6d831d
nullschema nullschema/1lfm25-8b-a1b-midrange llama.cpp extraction27.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c b3d59631a0a0785e
worldview worldview/2lfm25-8b-a1b-midrange llama.cpp extraction30.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 717aaa5a6a785c10
toolscale toolscale/1lfm25-8b-a1b-midrange llama.cpp extraction82.5%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 4727b9e997fb1a2e
toolcall toolcall/1lfm25-8b-a1b-midrange llama.cpp extraction1.7%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 3b306a765082433d
synthesis synthesis/2qwen3-8b-midrange llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 07092184bce46ed0
nullschema nullschema/1qwen3-8b-midrange llama.cpp extraction34.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 5fdc1502f911a148
worldview worldview/2qwen3-8b-midrange llama.cpp extraction45.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 39d8f0df5c0c0e6f
toolscale toolscale/1qwen3-8b-midrange llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c d27e6f32d963b586
toolcall toolcall/1qwen3-8b-midrange llama.cpp extraction0.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c c341d6e4665581b5
nullschema nullschema/1qwen38-vendor-qwen-think llama.cpp vendor-qwen-think56.2% voidCoverage 16/16g d7448e0f5fc3f9af d 3b91280bb40372ef c e25e7e28827fd589
worldview worldview/2qwen38-vendor-qwen-think llama.cpp vendor-qwen-think77.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c e78fad9e97d45282
synthesis synthesis/2qwen38-gsqrco-iq3s llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 92ea3b5723445268
nullschema nullschema/1qwen38-gsqrco-iq3s llama.cpp extraction62.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 62838af8410253c2
worldview worldview/2qwen38-gsqrco-iq3s llama.cpp extraction77.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 624bd85753786f4e
toolscale toolscale/1qwen38-gsqrco-iq3s llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 3177b556489558c1
toolcall toolcall/1qwen38-gsqrco-iq3s llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c d08352bc157ff4d3
synthesis synthesis/2bonsai-27b-ternary llama.cpp extraction66.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c ef7f6527771769a7
nullschema nullschema/1bonsai-27b-ternary llama.cpp extraction27.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 204142d8636f4fc2
worldview worldview/2bonsai-27b-ternary llama.cpp extraction42.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 121caa6718989cf9
toolscale toolscale/1bonsai-27b-ternary llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 83edc9129d9d77cd
toolcall toolcall/1bonsai-27b-ternary llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c b3a241de7392ea56
toolscale toolscale/1k2-36b-np1 llama.cpp extraction87.5%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 6e72c541e7fd804e
toolcall toolcall/1k2-36b-np1 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c f4f31defced2ec18
synthesis synthesis/2qwen38-27b-np1 llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 2867e80b35de0e49
nullschema nullschema/1qwen38-27b-np1 llama.cpp extraction76.4%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 458eb55eee2cc3a3
worldview worldview/2qwen38-27b-np1 llama.cpp extraction77.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c cf60f4eb1ae532f8
toolscale toolscale/1qwen38-27b-np1 llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 2b5ec5e83a2c8961
toolcall toolcall/1qwen38-27b-np1 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 7019f408382de810
nullschema nullschema/1k2-36b-vendor-k2b llama.cpp vendor-k2b57.7%Coverage 98.6% coverageg d7448e0f5fc3f9af d 3b91280bb40372ef c 62630cca2615a18f
synthesis synthesis/2k2-36b-vendor-k2b llama.cpp vendor-k2b90.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 19ff3eed9d7b1a47
worldview worldview/2k2-36b-vendor-k2b llama.cpp vendor-k2b82.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c d16f8472c1b1d87d
synthesis synthesis/2k2-36b-vendor-k2 llama.cpp vendor-k293.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c b7e765413a3dd2f2
nullschema nullschema/1k2-36b-vendor-k2 llama.cpp vendor-k248.6%Coverage 97.2% coverageg d7448e0f5fc3f9af d 3b91280bb40372ef c 80a8857ae078d242
worldview worldview/2k2-36b-vendor-k2 llama.cpp vendor-k277.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c e6f5010ab18f79e7
synthesis synthesis/2k2-36b-extraction llama.cpp extraction75.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c b92607640a462939
nullschema nullschema/1k2-36b-extraction llama.cpp extraction54.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 5e938f4fe9784ee9
worldview worldview/2k2-36b-extraction llama.cpp extraction52.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 9cc504b6f261a5b3
chainwalk chainwalk/3k2-36b-ctx128k-q8 llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c b3aa4c4b6efce314
worldview worldview/2k2-36b-ctx128k-q8 llama.cpp extraction65.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 082b24620003bf78
synthesis synthesis/2k2-36b-ctx32k llama.cpp extraction70.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 7873f3430a2f042e
toolscale toolscale/1k2-36b-ctx32k llama.cpp extraction90.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c b2321bc042bae630
toolcall toolcall/1k2-36b-ctx32k llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 3b6cfef3b606fe07
chainwalk chainwalk/3k2-36b-ctx32k llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c b3aa4c4b6efce314
screening screening/1k2-36b-ctx32k llama.cpp extraction90.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c af55e22558068ab0
nullschema nullschema/1k2-36b-ctx32k llama.cpp extraction54.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 345de57f0f3a686f
worldview worldview/2k2-36b-ctx32k llama.cpp extraction65.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 082b24620003bf78
worldview worldview/2k2-36b-a4b-q4 llama.cpp extraction78.9% voidCoverage 47.5% coverageg 9acfba1b704feb28 d e97835de9e6fc227 c 082b24620003bf78
synthesis synthesis/2k2-36b-a4b-q4 llama.cpp extraction73.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 7873f3430a2f042e
toolscale toolscale/1k2-36b-a4b-q4 llama.cpp extraction90.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c b2321bc042bae630
toolcall toolcall/1k2-36b-a4b-q4 llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 3b6cfef3b606fe07
worldview worldview/2k2-7b llama.cpp extraction30.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c e229d76d7d696757
worldview worldview/2k2-3.7b llama.cpp extraction50.0%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 13dc41ad1b2821d9
worldview worldview/2gemma4-12b llama.cpp extraction97.5%Coverage 40/40g 9acfba1b704feb28 d e97835de9e6fc227 c 585e7383f726d519
worldview worldview/1reap-60b mlx_lm extraction47.5%Coverage 40/40g d511b635ce7bbfd3 d d7ae9cbb38fafbe9 c f6b9a462ff423a87
synthesis synthesis/2gemma4-12b llama.cpp extraction83.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c dea865f1d077f0d8
synthesis synthesis/2k2-7b llama.cpp extraction75.0%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c d3f015ffb0c32d4d
synthesis synthesis/2k2-3.7b llama.cpp extraction58.3%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c be9a405fa957d750
synthesis synthesis/2reap-60b mlx_lm extraction81.7%Coverage 60/60g 4156dd3c13361e8b d 267d1954e225c5b6 c 6e88592c339b65ea
synthesis synthesis/1gemma4-12b llama.cpp extraction80.0%Coverage 60/60g 27df090db4445c2c d 760936a5348b2ffd c 9749137aeec83ad6
synthesis synthesis/1k2-7b llama.cpp extraction75.0%Coverage 60/60g 27df090db4445c2c d 760936a5348b2ffd c 58c73d3c88ebea78
synthesis synthesis/1k2-3.7b llama.cpp extraction86.7%Coverage 60/60g 27df090db4445c2c d 760936a5348b2ffd c cc272a124c5d1351
synthesis synthesis/1qwen3-next-80b-reap-mlx-mixed36 mlx_lm extraction100.0%Coverage 60/60g 27df090db4445c2c d 760936a5348b2ffd c 93c5f1d3b8c8e17c
chainwalk chainwalk/3k2-horizon-3.7b llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 497e8e74501f9a9e
nullschema nullschema/1k2-horizon-3.7b llama.cpp extraction43.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c c1cd134c67de1621
screening screening/1k2-horizon-3.7b llama.cpp extraction100.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 5d3e0bfbb30467e0
tools tools/1k2-horizon-3.7b llama.cpp extraction86.7%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 4b2e40f5222213ec
toolscale toolscale/1k2-horizon-3.7b llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 4535019c467a8444
toolcall toolcall/1k2-horizon-3.7b llama.cpp extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c f848b35c571d582e
tools tools/1k2-horizon-7b llama.cpp extraction100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c d1b260004a9b283e
chainwalk chainwalk/3k2-horizon-7b llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c ce69ff4b8656191d
toolscale toolscale/1k2-horizon-7b llama.cpp extraction100.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 62b1c628753b16b6
grounding grounding/1k2-horizon-7b llama.cpp extraction55.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 30ddb902b06a0c00
nullschema nullschema/1k2-horizon-7b llama.cpp extraction34.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 54f9df9571bc16e8
screening screening/1k2-horizon-7b llama.cpp extraction90.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c e92045e24be569c7
contract contract/2k2-horizon-7b llama.cpp extraction22.9%Coverage 48/48g cc2d23c5bae45474 d 1fe53d30cb5480f6 c e127d0663cbe7e98
toolcall toolcall/1k2-horizon-7b llama.cpp extraction95.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 93ba527fea40c309
screening screening/1granite-guardian-4.1-8b-mlx-bf16 mlx_lm extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 081a28b5708d9572
screening screening/1granite-guardian-4.1-8b-mlx-q8 mlx_lm extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 764b495364f3e7ff
screening screening/1granite-guardian omlx extraction36.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 9de3a458bda36d8c
screening screening/1granite-guardian-4.1-8b-mlx-mixed36 mlx_lm extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 1e2bab38313786db
screening screening/1guardian-mixed36-mlxlm mlx_lm extractionn/a voidCoverage 0% coverageg 0661b49a10be2d2a d 64c97ca2269c29af c b56b3c67a8112f7f
screening screening/1granite-guardian llama.cpp extraction100.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c fe46fd4c1301cd33
screening screening/1granite-guardian llama.cpp extraction100.0% voidCoverage 41/41g 0661b49a10be2d2a d 64c97ca2269c29af c fe46fd4c1301cd33
screening screening/1granite-guardian-gguf-q4km llama.cpp extraction100.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 6af9b4aa83361ef8
screening screening/1granite-guardian omlx extraction37.3% voidCoverage 59/59g 0661b49a10be2d2a d 64c97ca2269c29af c 9de3a458bda36d8c
screening screening/1granite-guardian-gguf-q4km llama.cpp extraction100.0% voidCoverage 36/36g 0661b49a10be2d2a d 64c97ca2269c29af c 6af9b4aa83361ef8
screening screening/1granite-guardian omlx extraction40.0% voidCoverage 5/5g 0661b49a10be2d2a d 64c97ca2269c29af c 9de3a458bda36d8c
screening screening/1granite-guardian-gguf-q4km llama.cpp extraction100.0% voidCoverage 33/33g 0661b49a10be2d2a d 64c97ca2269c29af c 6af9b4aa83361ef8
screening screening/1granite-guardian-mlx-mixed36 omlx extractionn/a voidCoverage 0% coverageg 0661b49a10be2d2a d 64c97ca2269c29af c d9be4d5a49d0f4d5
toolcall toolcall/1qwen3-next-80b-reap omlx extraction35.0% voidCoverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c e9925436d57909a1
toolcall toolcall/1gemma4-e2b omlx extraction0.0% voidCoverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 022b5819dd82ff96
toolcall toolcall/1granite-4.2-30b omlx extraction20.0% voidCoverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c bf1109eb4e492a3d
toolcall toolcall/1qwen3.6-35b omlx extraction65.0% voidCoverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 8e5c69081ab588b0
toolcall toolcall/1qwen3.8-27b omlx extraction100.0% voidCoverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c ac125d630f2b70bc
toolcall toolcall/1mlx-community-Qwen3.6-35B-A3B-4bit rapid-mlx extraction60.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 4a82e7591c905bf0
toolcall toolcall/1qwen3-next-80b-reap-mlx-mixed36 rapid-mlx extraction78.3%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 8352ab9b62086e37
toolcall toolcall/1qwen3.8-27b-mlx-4bit rapid-mlx extraction100.0%Coverage 60/60g 804e8fc9457649d8 d 10af88c170513155 c 8886d10bc276895c
contract contract/2mlx-community-Qwen3.6-35B-A3B-4bit rapid-mlx extraction27.1%Coverage 48/48g 486601ba23034dd6 d 1fe53d30cb5480f6 c 4fc6c3fdb9c9b02c
contract contract/2q36-toolparser rapid-mlx extractionn/a voidCoverage 0% coverageg 486601ba23034dd6 d 1fe53d30cb5480f6 c 2225a59526198b68
screening screening/1mlx-community-Qwen3.6-35B-A3B-4bit rapid-mlx extraction90.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 4ed45f89d9eca30b
screening screening/1qwen3.8-27b-mlx-4bit rapid-mlx extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 1b2a598c6ef11638
contract contract/2qwen3.8-27b-mlx-4bit rapid-mlx extraction75.0%Coverage 48/48g 486601ba23034dd6 d 1fe53d30cb5480f6 c 2750a525d9018a88
contract contract/2flashnext-4.27bpw-m64 llama.cpp extractionn/a voidCoverage 0% coverageg 486601ba23034dd6 d 1fe53d30cb5480f6 c 92b28699e44d08e7
contract contract/2qwen3-next-80b-reap llama.cpp extractionn/a voidCoverage 0% coverageg 486601ba23034dd6 d 1fe53d30cb5480f6 c cbb517440b51fc70
contract contract/2qwen3-next-80b-reap-mlx-mixed36 rapid-mlx extraction54.2%Coverage 48/48g 486601ba23034dd6 d 1fe53d30cb5480f6 c e059879c3195b24c
contract contract/1qwen3-next-80b-reap-mlx-q4 mlx_lm extraction47.2%Coverage 36/36g 2dd146852463f032 d ab4bb33923bd0ed6 c 14fac1620eb065e0
screening screening/1qwen3-next-80b-reap-mlx-q4 mlx_lm extraction100.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 72e18625ddbdaf10
screening screening/1qwen3-next-80b-reap-mlx-mixed36 rapid-mlx extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c c37510db6b5ff9cf
contract contract/1reapq4-mlx_lm mlx_lm extractionn/a voidCoverage 0% coverageg 2dd146852463f032 d ab4bb33923bd0ed6 c 224bdfd4a5d2e672
screening screening/1reapq4-mlx_lm mlx_lm extractionn/a voidCoverage 0% coverageg 0661b49a10be2d2a d 64c97ca2269c29af c 2f64770a4290f344
contract contract/1reapmix36-rapid-mlx rapid-mlx extractionn/a voidCoverage 0% coverageg 2dd146852463f032 d ab4bb33923bd0ed6 c 46d4a279b80fca4d
screening screening/1reapmix36-rapid-mlx rapid-mlx extractionn/a voidCoverage 0% coverageg 0661b49a10be2d2a d 64c97ca2269c29af c 8077358e07c7af36
contract contract/1reapq4-rapid-mlx rapid-mlx extractionn/a voidCoverage 0% coverageg 2dd146852463f032 d ab4bb33923bd0ed6 c 788a11f836a3345b
screening screening/1reapq4-rapid-mlx rapid-mlx extractionn/a voidCoverage 0% coverageg 0661b49a10be2d2a d 64c97ca2269c29af c ba21b624ebdf49bd
contract contract/1granite-4.2-30b llama.cpp extraction66.7%Coverage 36/36g 2dd146852463f032 d ab4bb33923bd0ed6 c cc8ca71e6e299aee
contract contract/1gemma4-12b llama.cpp extraction66.7%Coverage 36/36g 2dd146852463f032 d ab4bb33923bd0ed6 c 0d1de531e8b00c72
chainwalk chainwalk/3flashnext-4.27bpw-m64 llama.cpp agentic-fast0.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 17c29ba75eeadc76
screening screening/1flashnext-4.27bpw-m64 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 028bac48a65abdd9
nullschema nullschema/1flashnext-4.27bpw-m64 llama.cpp extraction58.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c a1f73dd74c73dbc6
grounding grounding/1flashnext-4.27bpw-m64 llama.cpp extraction81.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 2867d56b04f8bbb4
chainwalk chainwalk/3qwen3-next-80b-reap-ag llama.cpp agentic-fast50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c ccae884465959abc
toolscale toolscale/1qwen3-next-80b-reap-ag llama.cpp agentic-fast90.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 2b03adebd9317650
tools tools/1qwen3-next-80b-reap-ag llama.cpp agentic-fast100.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 3907525a897ced43
nullschema nullschema/1qwen3-next-80b-reap llama.cpp extraction97.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c cc55008d40e60573
grounding grounding/1qwen3-next-80b-reap llama.cpp extraction87.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 7e1bbf58551d87ef
screening screening/1qwen3-next-80b-reap llama.cpp extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 81efb9eaae394534
screening screening/1glm-4.7-flash-mlx-mixed_4_6 mlx_lm extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c c1b73af4ac52ca1f
screening screening/1glm-4.7-flash-mlx-mixed_3_6 mlx_lm extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 99894ae40745ad56
grounding grounding/1muse-glimmer-30b llama.cpp extraction85.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c d9d25da365e04cc8
screening screening/1muse-glimmer-30b llama.cpp extraction86.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 63338a6fdcbdcddd
screening screening/1laguna-xs2-mlx-q4 mlx_lm extractionn/a voidCoverage 0% coverageg 0661b49a10be2d2a d 64c97ca2269c29af c 3c17fa93e654c0b4
nullschema nullschema/1glm-4.7-flash-mlx-q4 mlx_lm extraction40.3% voidCoverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c b669c85bfa68b0b5
grounding grounding/1glm-4.7-flash-mlx-q4 mlx_lm extraction53.6% voidCoverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 2432689cc6751eea
screening screening/1glm-4.7-flash-mlx-q4 mlx_lm extraction96.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c c80ceb72756eafac
nullschema nullschema/1glm-4.7-flash-mlx-q8 mlx_lm extraction59.7% voidCoverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 97ef09a912feb8be
grounding grounding/1glm-4.7-flash-mlx-q8 mlx_lm extraction76.8% voidCoverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 6d0a345ab4ecfa54
screening screening/1glm-4.7-flash-mlx-q8 mlx_lm extraction96.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c b4008e66dcadf553
nullschema nullschema/1qwen3-8b-mlx-8bit mlx_lm extraction56.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c b9aaca6eee78b218
grounding grounding/1qwen3-8b-mlx-8bit mlx_lm extraction74.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c e34a89daf5181d58
screening screening/1qwen3-8b-mlx-8bit mlx_lm extraction96.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 02039a8c6b78a23b
nullschema nullschema/1nemotron-lightning-30b-mlx-4bit mlx_lm extraction9.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 023ab1ea82fce0de
grounding grounding/1nemotron-lightning-30b-mlx-4bit mlx_lm extraction56.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 15a43d7b2024f68c
screening screening/1nemotron-lightning-30b-mlx-4bit mlx_lm extraction90.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 813491de8103980c
nullschema nullschema/1granite-4.2-30b-gguf llama.cpp extraction90.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c ad5177ffa2d55fa5
grounding grounding/1granite-4.2-30b-gguf llama.cpp extraction84.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c e353f603986dcdcd
screening screening/1granite-4.2-30b-gguf llama.cpp extraction98.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 001987450ecbaa8a
nullschema nullschema/1ornith-1.5-35b-mlx-4bit mlx_lm extraction62.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 1fb1a5ad2ee784e3
grounding grounding/1ornith-1.5-35b-mlx-4bit mlx_lm extraction67.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c a438aff9707ec8a1
screening screening/1ornith-1.5-35b-mlx-4bit mlx_lm extraction88.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 4f3b9812298e963e
nullschema nullschema/1granite-4.2-30b-mlx-q4 mlx_lm extraction100.0% voidCoverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 48b3d526e6d6e6a3
VOID-degenerate-nullschema nullschema/1granite-4.2-30b-mlx-q4 mlx_lm extraction100.0% voidCoverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 48b3d526e6d6e6a3
grounding grounding/1granite-4.2-30b-mlx-q4 mlx_lm extraction86.4% voidCoverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c db0a9174aa469c68
screening screening/1granite-4.2-30b-mlx-q4 mlx_lm extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c f71ccbbb16cd8099
shapegen shapegen/2qwen3-coder-30b llama.cpp extraction0.0%Coverage 12/12g f23222ad8df22fd8 d fc27d4e1a999c675 c cbf6852b8edc0984
sitebuild sitebuild/1qwen3-coder-30b llama.cpp agentic-fast0.0% voidCoverage 1/1g b2c603d6c0d6afb3 d 8b54725d2778dc60 c 0cac83c8fbc03eec
tagger_nullable tagger-nullable/1tagger-e2b-v1-dl-tagger-e2b-v1 mlx_lm extraction75.0%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c b0015c0679b3c9f8
tagger tagger/2tagger-e2b-v1-dl-tagger-e2b-v1 mlx_lm extraction74.2%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 26b77eacaf1fc207
tagger_nullable tagger-nullable/1gemma4-e2b-mlx-4bit mlx_lm extraction73.8%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c d40c89a4c0ded16e
tagger tagger/2gemma4-e2b-mlx-4bit mlx_lm extraction85.8%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 1623b2cdd0ccbfaf
nullschema nullschema/1qwen3-next-80b-mlx-4bit mlx_lm extraction95.8% voidCoverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c d1b64a4d48cff9ef
screening screening/1qwen3-next-80b-mlx-4bit mlx_lm extraction96.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 22413c6aa05bedb0
grounding grounding/1qwen3-next-80b-mlx-4bit mlx_lm extraction87.2% voidCoverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 0c539d39afff5bd1
grounding grounding/1qwen3.8-27b-noreason llama.cpp extraction80.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 4a3b4b7c53219329
grounding grounding/1qwen3.8-27b-mlx-4bit mlx_lm extraction92.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 42a3cf291ff302bb
chainwalk chainwalk/3qwen-agentworld-35b llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c c4305e58879e2b40
chainwalk chainwalk/3qwen3-next-80b-ag llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 3a96fdcc21129ae7
chainwalk chainwalk/3qwen3.6-35b-ag llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c ea9f2c30d07a149b
chainwalk chainwalk/3qwen3.6-35b-ag llama.cpp extraction50.0% voidCoverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c ea9f2c30d07a149b
toolscale toolscale/1qwen3.6-35b-ag llama.cpp extraction80.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c f995b25bd0b360cf
tools tools/1qwen3.6-35b-ag llama.cpp extraction93.3%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c a8b21b992370e7c6
chainwalk chainwalk/3qwen3-next-80b-ag llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c 3a96fdcc21129ae7
toolscale toolscale/1qwen3-next-80b-ag llama.cpp extraction95.0%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c 18366d2ac6e48e7c
tools tools/1qwen3-next-80b-ag llama.cpp extraction93.3%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c 91dcd2c830490d0e
chainwalk chainwalk/3qwen-agentworld-35b llama.cpp extraction50.0%Coverage 2/2g f4bfcf6f040b0978 d 129a9e74408c8919 c c4305e58879e2b40
toolscale toolscale/1qwen-agentworld-35b llama.cpp extraction72.5%Coverage 40/40g a2bc3751abfa28e0 d b24abb108c409178 c edcc1fd1f9bd35f0
tools tools/1qwen-agentworld-35b llama.cpp extraction80.0%Coverage 15/15g 2763675a5c3da564 d 33a44812f0ad4c39 c e56fd9e9b9103b0a
screening screening/1granite-guardian-3.1-2b llama.cpp extraction0.0% voidCoverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c c3bc4d1194d50b4b
nullschema nullschema/1qwen38-flash-next llama.cpp extraction50.0%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c d248e89da830ce31
screening screening/1qwen38-flash-next llama.cpp extraction96.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c dd9b4d06d5abee09
grounding grounding/1qwen38-flash-next llama.cpp extraction79.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 9148fcdc2bc14fe4
nullschema nullschema/1nemotron-lightning-30b llama.cpp extraction15.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 8f3225e9b30dcc6b
screening screening/1nemotron-lightning-30b llama.cpp extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c bd58c270aedc9f3d
grounding grounding/1nemotron-lightning-30b llama.cpp extraction60.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c f734b41b3caabfdd
nullschema nullschema/1qwen3-next-80b llama.cpp extraction97.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 5913b00d3e8c6b3d
screening screening/1qwen3-next-80b llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c ede1a2072fbaea6c
grounding grounding/1qwen3-next-80b llama.cpp extraction88.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c f34a40087b318cc6
tagger_nullable tagger-nullable/1qwen3-next-80b-tag llama.cpp extraction90.4%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c c033312a94ce8bbe
tagger tagger/2qwen3-next-80b-tag llama.cpp extraction93.3%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c bb9d18596364b9f4
tagger_nullable tagger-nullable/1qwen3.6-35b-tag llama.cpp extraction92.5%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 2f1c411d3dd00d20
tagger tagger/2qwen3.6-35b-tag llama.cpp extraction90.8%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 336518ff1b1d0973
tagger_nullable tagger-nullable/1gemma4-12b-b10618 llama.cpp extraction97.9%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c e947bd43665a9bdb
tagger tagger/2gemma4-12b-b10618 llama.cpp extraction93.8%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c c947e8a080eadfe2
screening screening/1granite-guardian-4.1-8b llama.cpp extraction100.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 9ba7cd65a17e6418
nullschema nullschema/1lfm2.5-8b-a1b llama.cpp extraction27.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 00612fdd4a39faf4
grounding grounding/1lfm2.5-8b-a1b llama.cpp extraction47.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 90be8dffe935ab41
screening screening/1lfm2.5-8b-a1b llama.cpp extraction73.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 2cbf094131c24516
nullschema nullschema/1qwen3-next-80b-coder llama.cpp extraction56.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 74810b0cff4603b3
screening screening/1qwen3-next-80b-coder llama.cpp extraction90.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c e7b3303ff47f6803
grounding grounding/1qwen3-next-80b-coder llama.cpp extraction76.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 140e26c8be639be7
nullschema nullschema/1qwen3.6-35b llama.cpp extraction61.1%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 7a652a6a84f42ef5
screening screening/1qwen3.6-35b llama.cpp extraction86.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 4246ff56e771b097
grounding grounding/1qwen3.6-35b llama.cpp extraction82.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 279f3a4bd1ff0c7e
grounding grounding/1qwen3.6-35b llama.cpp extraction86.3% voidCoverage 73/73g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 279f3a4bd1ff0c7e
grounding grounding/1qwen3-next-80b-thinking llama.cpp study-default100.0% voidCoverage 0.8% coverageg 30b933dcc6a7e2d4 d dcb396faf8b83b0a c cb1f0b0a42b6ac48
nullschema nullschema/1qwen3-next-80b-thinking llama.cpp extraction58.3%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 8f8aa14bba9f91c3
screening screening/1qwen3-next-80b-thinking llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 4320932025e05e5d
grounding grounding/1qwen3-next-80b-thinking llama.cpp extraction65.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c e644b21297e6198b
nullschema nullschema/1qwen3-next-80b llama.cpp extraction97.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 5913b00d3e8c6b3d
screening screening/1qwen3-next-80b llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c ede1a2072fbaea6c
grounding grounding/1qwen3-next-80b llama.cpp extraction88.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c f34a40087b318cc6
nullschema nullschema/1qwen3-next-80b-a3b-iq4xs llama.cpp extraction97.2%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 860f8678b21dfb1f
screening screening/1qwen3-next-80b-a3b-iq4xs llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c c6cb8e26e5eec1ec
grounding grounding/1qwen3-next-80b-a3b-iq4xs llama.cpp extraction88.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c e3b83bfa70c1a417
grounding grounding/1qwen3-next-80b-a3b-iq4xs-smoke llama.cpp extraction73.9% voidCoverage 92% coverageg 30b933dcc6a7e2d4 d dcb396faf8b83b0a c e6920f17163814d7
nullschema nullschema/1glm-4.7-flash-q8 llama.cpp extraction48.6%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 3d836e2b73f9cd77
screening screening/1glm-4.7-flash-q8 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c b7d076058290ae5f
grounding grounding/1glm-4.7-flash-q8 llama.cpp extraction76.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c f7116f1eb6a7f001
nullschema nullschema/1glm-4.7-flash-q8 llama.cpp extractionn/a voidCoverage 0% coverageg d7448e0f5fc3f9af d 3b91280bb40372ef c 3d836e2b73f9cd77
screening screening/1glm-4.7-flash-q8 llama.cpp extractionn/a voidCoverage 0% coverageg 0661b49a10be2d2a d 64c97ca2269c29af c b7d076058290ae5f
nullschema nullschema/1qwen3.5-122b-a10b-iq3 llama.cpp extraction45.8%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 5e7cd3243c74ce98
screening screening/1qwen3.5-122b-a10b-iq3 llama.cpp extraction88.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 45dab17de7814804
grounding grounding/1qwen3.5-122b-a10b-iq3 llama.cpp extraction80.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c b3059734754cd76c
nullschema nullschema/1agents-a1-35b-b10618 llama.cpp extraction38.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 9a9ee4f1853f385a
screening screening/1agents-a1-35b-b10618 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c fe6603a03488c606
grounding grounding/1agents-a1-35b-b10618 llama.cpp extraction71.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c b6d15c86cef9c0f6
nullschema nullschema/1qwen3-coder-30b-b10618 llama.cpp extraction37.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c ffb75533d79bc27d
screening screening/1qwen3-coder-30b-b10618 llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 01f8fd00f78cb299
grounding grounding/1qwen3-coder-30b-b10618 llama.cpp extraction76.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 85522d05c3671c16
nullschema nullschema/1qwen3.5-9b-q8-b10618 llama.cpp extraction30.6%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 989a2b1d084c0c53
screening screening/1qwen3.5-9b-q8-b10618 llama.cpp extraction96.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 6c6da4df0333507c
grounding grounding/1qwen3.5-9b-q8-b10618 llama.cpp extraction72.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 9e54265c33ff0f82
nullschema nullschema/1gemma4-12b-b10618 llama.cpp extraction88.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 2c03d6ebad98fccc
screening screening/1gemma4-12b-b10618 llama.cpp extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 79323c88f8e515df
grounding grounding/1gemma4-12b-b10618 llama.cpp extraction84.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c ba97a71cfa0468a7
nullschema nullschema/1gemma4-e2b-b10618 llama.cpp extraction12.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 38b4b96050d74d17
screening screening/1gemma4-e2b-b10618 llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 80f542cbf88cfba7
grounding grounding/1gemma4-e2b-b10618 llama.cpp extraction61.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c b18ad353c84d839e
nullschema nullschema/1gemma4-e4b-b10618 llama.cpp extraction13.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 84794dc91d0ce3b4
screening screening/1gemma4-e4b-b10618 llama.cpp extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c b75251f174426f4d
grounding grounding/1gemma4-e4b-b10618 llama.cpp extraction62.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c dc6988d619b065bf
nullschema nullschema/1ornith-1.5-35b-q6-b10618 llama.cpp extraction34.7%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 355fd8b860ca54b1
screening screening/1ornith-1.5-35b-q6-b10618 llama.cpp extraction95.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c ce44f5f606a0131e
grounding grounding/1ornith-1.5-35b-q6-b10618 llama.cpp extraction72.8%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c bc2f7777482c6fb6
nullschema nullschema/1devstral-small-2-24b-b10618 llama.cpp extraction62.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c 836a74186efb247a
screening screening/1devstral-small-2-24b-b10618 llama.cpp extraction100.0%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c d8a2ef703d82163d
grounding grounding/1devstral-small-2-24b-b10618 llama.cpp extraction79.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c ed588aa2abdb41ef
screening screening/1gemma4-12b llama.cpp extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c b7d605769821c0d8
nullschema nullschema/1gemma4-12b llama.cpp extraction88.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c ec3358e6aa5e1088
screening screening/1gemma4-e4b llama.cpp extraction93.3%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 824bef7dab174615
nullschema nullschema/1gemma4-e4b llama.cpp extraction13.9%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c addbcab5facddd0b
screening screening/1gemma4-e2b llama.cpp extraction91.7%Coverage 60/60g 0661b49a10be2d2a d 64c97ca2269c29af c 905fc96bd8c3e874
nullschema nullschema/1gemma4-e2b llama.cpp extraction12.5%Coverage 72/72g d7448e0f5fc3f9af d 3b91280bb40372ef c e268490afdec64ce
tagger_nullable tagger-nullable/1gemma4-12b-ab llama.cpp extraction97.9%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 8bf8e2bee9ac7333
tagger tagger/2gemma4-12b-ab llama.cpp extraction93.8%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 3c8ef5d04a11b585
tagger_nullable tagger-nullable/1gemma4-e2b-ab llama.cpp extraction90.4%Coverage 240/240g 04e875431e26c4b5 d db3addf23a2bc4ce c 23e712860bee6d0e
tagger tagger/2gemma4-e2b-ab llama.cpp extraction91.2%Coverage 240/240g 63e04d139a0e5ed3 d 7ef587bf33e5788a c 4619919ac18affd5
grounding grounding/1qwen3-coder-30b llama.cpp extraction77.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c ade06bde17b71832
grounding grounding/1qwen3.8-27b llama.cpp extraction80.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 1d989a6b00eb212c
grounding grounding/1agents-a1-35b llama.cpp extraction71.2%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 3cb35299a047efb8
grounding grounding/1qwen3.5-9b-q8 llama.cpp extraction72.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c bd21ebb86ee6f863
grounding grounding/1gemma4-12b llama.cpp extraction84.0%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 31241ed46ed4d6a5
grounding grounding/1gemma4-e4b llama.cpp extraction62.4%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c 267624907124ed7e
grounding grounding/1gemma4-e2b llama.cpp extraction61.6%Coverage 125/125g 30b933dcc6a7e2d4 d dcb396faf8b83b0a c dee67653e2de45c6
tagger tagger/2qwen3.8-iq1s llama.cpp study-nothink31.6% voidCoverage 44.2% coverageg 956e03bbffa79aa4 d 7ef587bf33e5788a c f6700e17f70a9c2a
epistemics epistemics/4qwen3.8-iq1s llama.cpp study-nothink77.1%Coverage 72.9% coverageg 2a2f7e4e90d6dd63 d 4bce453fd376553c c 4bfecdedff811923
tools tools/1qwen3.8-iq1s llama.cpp agentic-fast76.9%Coverage 86.7% coverageg d6a32c09e57bea7b d 33a44812f0ad4c39 c 02e356bf0012eaf2
enrichment enrichment/3qwen3.8-iq1s llama.cpp study-nothink60.5%Coverage 79.6% coverageg 7d661b93852b5ffd d 75361864baa727ab c 11d22e9108889bce
chainwalk chainwalk/3qwen3.8-iq1s llama.cpp agentic-fast0.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c e2e5d225bbfe60d6
chainwalk chainwalk/3gemma4-12b-r3 llama.cpp agentic-fast50.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c 2cbb0e64811ac557
chainwalk chainwalk/3gemma4-e4b-r3 llama.cpp agentic-fast0.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c 7e97acd7556a3909
chainwalk chainwalk/3gemma4-e2b-r3 llama.cpp agentic-fast0.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c 24f4c1683cfb7a85
chainwalk chainwalk/3gemma4-12b-r2 llama.cpp agentic-fast50.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c e1a3cdf8d371cae6
chainwalk chainwalk/3gemma4-e4b-r2 llama.cpp agentic-fast0.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c ec61f0c3716fc3f8
chainwalk chainwalk/3gemma4-e2b-r2 llama.cpp agentic-fast0.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c 277579259f233168
toolscale toolscale/1agents-a1-35b llama.cpp agentic-fast100.0%Coverage 40/40g 1bae1f3e1b606eae d b24abb108c409178 c f55610aeacd8ae8c
chainwalk chainwalk/3gemma4-12b-r1 llama.cpp agentic-fast50.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c 97645e44ce72b737
toolscale toolscale/1gemma4-12b llama.cpp agentic-fast90.0%Coverage 40/40g 1bae1f3e1b606eae d b24abb108c409178 c 42412bb32a456ced
toolscale toolscale/1gemma4-e4b llama.cpp agentic-fast100.0%Coverage 40/40g 1bae1f3e1b606eae d b24abb108c409178 c 64d002411f012d3c
toolscale toolscale/1gemma4-e2b llama.cpp agentic-fast100.0%Coverage 40/40g 1bae1f3e1b606eae d b24abb108c409178 c 3860ab39a73cff08
chainwalk chainwalk/3gemma4-e4b-r1 llama.cpp agentic-fast0.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c 7ef98c4a8fe4369b
chainwalk chainwalk/3gemma4-e2b-r1 llama.cpp agentic-fast0.0%Coverage 2/2g d31d3c4c6285617f d 129a9e74408c8919 c 92c0ac5ca29b9cdc
epistemics epistemics/4qwen3.8-27b llama.cpp study-nothink100.0%Coverage 48/48g 2a2f7e4e90d6dd63 d 4bce453fd376553c c 541a385213b500cb
tools tools/1qwen3.8-27b llama.cpp agentic-fast100.0%Coverage 15/15g d6a32c09e57bea7b d 33a44812f0ad4c39 c 35475e0c17bfc385
enrichment enrichment/3gemma4-12b llama.cpp study-nothink77.8%Coverage 54/54g 7d661b93852b5ffd d 75361864baa727ab c 5367be2a63165de3
enrichment enrichment/3gemma4-e4b llama.cpp study-nothink70.4%Coverage 54/54g 7d661b93852b5ffd d 75361864baa727ab c 60945fac08964a72
enrichment enrichment/3gemma4-e2b llama.cpp study-nothink75.9%Coverage 54/54g 7d661b93852b5ffd d 75361864baa727ab c 94079faf08beb84b
epistemics epistemics/4agents-a1-35b llama.cpp study-nothink100.0%Coverage 48/48g 2a2f7e4e90d6dd63 d 4bce453fd376553c c 2101ffcef80989dc
tools tools/1agents-a1-35b llama.cpp agentic-fast100.0%Coverage 15/15g d6a32c09e57bea7b d 33a44812f0ad4c39 c 555c28b5386478f6
epistemics epistemics/4gemma4-12b llama.cpp study-nothink100.0%Coverage 48/48g 2a2f7e4e90d6dd63 d 4bce453fd376553c c b6ac24eba6a83a5d
epistemics epistemics/4gemma4-e4b llama.cpp study-nothink97.9%Coverage 48/48g 2a2f7e4e90d6dd63 d 4bce453fd376553c c 37147fa352af2d7a
epistemics epistemics/4gemma4-e2b llama.cpp study-nothink68.8%Coverage 48/48g 2a2f7e4e90d6dd63 d 4bce453fd376553c c e5abf1e2bbe8c2c8
chainwalk chainwalk/2gemma4-12b llama.cpp agentic-fast50.0%Coverage 2/2g 48abd22e0117fdda d 129a9e74408c8919 c b4834ee71de7e834
chainwalk chainwalk/2gemma4-e4b llama.cpp agentic-fast0.0%Coverage 2/2g 48abd22e0117fdda d 129a9e74408c8919 c 8e18e895c23f4485
chainwalk chainwalk/2gemma4-e2b llama.cpp agentic-fast0.0%Coverage 2/2g 48abd22e0117fdda d 129a9e74408c8919 c eef56b755300de59
tools tools/1gemma4-e2b llama.cpp agentic-fast100.0%Coverage 15/15g d6a32c09e57bea7b d 33a44812f0ad4c39 c 33f8efb324ad39aa
tools tools/1gemma4-12b llama.cpp agentic-fast86.7%Coverage 15/15g d6a32c09e57bea7b d 33a44812f0ad4c39 c b34ee497223231c8
tools tools/1gemma4-e4b llama.cpp agentic-fast100.0%Coverage 15/15g d6a32c09e57bea7b d 33a44812f0ad4c39 c 50b166f799c52d51
tagger tagger/2gemma4-12b llama.cpp study-nothink93.3%Coverage 80.4% coverageg 28983856e4bf2dd9 d 7ef587bf33e5788a c d10e58be49cb1d55
tagger tagger/2gemma4-e4b llama.cpp study-nothink90.8%Coverage 240/240g 28983856e4bf2dd9 d 7ef587bf33e5788a c 60e6c55236269057
tagger tagger/2gemma4-e2b llama.cpp study-nothink91.2%Coverage 240/240g 28983856e4bf2dd9 d 7ef587bf33e5788a c 72d758d807701d51
tagger tagger/1tagger-e2b-v1 mlx_lm extraction77.1%Coverage 240/240g e8205e9fa8fb5146 d 08de96244bc929b8 c —