Model← Models
Qwen3-Next-80B-A3B Instruct, REAP pruned
60B mixture of experts, ~3B active, IQ4_XS, 30.4 GB.
- Parameters
- 60B / ~3B active
- Quant
- IQ4_XS
- Size
- 30.4 GB
- Family
- qwen3-next
qwen3-next-80b-reap
Measured, not seatedMoEresidentllama.cpp
Model card ↗Outcome as of 20 September 2026, from the seat registry. The write-up below was made when the model was tested.
Why it was tested
Tested to find out whether expert pruning costs the extraction seat its abstention. The full model holds that seat at 40 GB, which is most of the memory budget on a 64 GB machine.
What came of it
It costs nothing measurable. nullschema 97.2 percent, the same 70 of 72 as the unpruned model with the same two failures. grounding within noise at 87.2 against 88.8, screening better at 98.3 against 91.7, and 41.4 tok/s against 37.7. Ten gigabytes smaller and slightly faster.
| Suite | Score | 95% interval | Passed | Runtime · profile | Recorded |
|---|---|---|---|---|---|
| contract | 0.0% | 0.0–7.4 | Passed 0/48 | llama.cppextraction | — |
| grounding | 87.2% | 80.2–92.0 | Passed 109/125 | llama.cppextraction | 2026-09-08 |
| nullschema | 97.2% | 90.4–99.2 | Passed 70/72 | llama.cppextraction | 2026-09-08 |
| screening | 98.3% | 91.1–99.7 | Passed 59/60 | llama.cppextraction | 2026-09-08 |
Runs marked unenforced were measured through mlx_lm on a suite that requires constrained decoding. mlx_lm accepts the schema and ignores it, so those numbers do not measure what the suite is designed to measure. They are kept and labelled rather than deleted.
Every run log entry for this model (5)
| Finished | Suite | Runtime · profile | Score | Coverage | Grader · dataset · config |
|---|---|---|---|---|---|
| toolcalltoolcall/1 | not recordedextraction | 35.0% | Coverage 60/60 | g 804e8fc9457649d8d 10af88c170513155c e9925436d57909a1 | |
| contractcontract/2 | not recordedextraction | n/a | Coverage 0% coverage | g 486601ba23034dd6d 1fe53d30cb5480f6c cbb517440b51fc70 | |
| nullschemanullschema/1 | not recordedextraction | 97.2% | Coverage 72/72 | g d7448e0f5fc3f9afd 3b91280bb40372efc cc55008d40e60573 | |
| groundinggrounding/1 | not recordedextraction | 87.2% | Coverage 125/125 | g 30b933dcc6a7e2d4d dcb396faf8b83b0ac 7e1bbf58551d87ef | |
| screeningscreening/1 | not recordedextraction | 98.3% | Coverage 60/60 | g 0661b49a10be2d2ad 64c97ca2269c29afc 81efb9eaae394534 |
- Qwen3-Next-80B-A3B InstructRemoved
80B · MoE · IQ4_XS · 40.0 GB
The incumbent. Chosen because 3B active parameters give MoE speed at 80B quality, and it fits at 40GB with room for a long context.
- Qwen3-Next-80B-A3B ThinkingRejected
80B · MoE · IQ4_XS · 40.0 GB
Tested to see whether reasoning improves extraction, on the assumption that thinking helps.