Qwen3.6-35B-A3B · file← Qwen3.6-35B-A3B
UD-Q5_K_M with MTP head
The first copy and the everyday model from 11 August. All the llama.cpp tests ran on it; still set up to load on demand. Unsloth GGUF.
- Size
- 27.1 GB
- Status
- On disk
- Scored runs
- 8
- Smallest Mac
- 48 GB estimate, 8K context
| Suite | Score | 95% interval | Passed | Runtime · profile | Recorded |
|---|---|---|---|---|---|
| chainwalkqwen3.6-35b-ag | 50.0% | 9.5–90.5 | Passed 1/2 | llama.cppextraction | 2026-09-07 |
| groundingqwen3.6-35b | 82.4% | 74.8–88.1 | Passed 103/125 | llama.cppextraction | 2026-08-27 |
| nullschemaqwen3.6-35b | 61.1% | 49.6–71.5 | Passed 44/72 | llama.cppextraction | 2026-08-27 |
| screeningqwen3.6-35b | 86.7% | 75.8–93.1 | Passed 52/60 | llama.cppextraction | 2026-08-27 |
| taggerqwen3.6-35b-tag | 90.8% | 86.5–93.9 | Passed 218/240 | llama.cppextraction | 2026-08-31 |
| tagger_nullableqwen3.6-35b-tag | 92.5% | 88.5–95.2 | Passed 222/240 | llama.cppextraction | 2026-08-31 |
| toolsqwen3.6-35b-ag | 93.3% | 70.2–98.8 | Passed 14/15 | llama.cppextraction | 2026-09-07 |
| toolscaleqwen3.6-35b-ag | 80.0% | 65.2–89.5 | Passed 32/40 | llama.cppextraction | 2026-09-07 |
Runs marked unenforced were measured on a suite that requires constrained decoding, on a server shown not to apply the schema: mlx_lm accepts it and ignores it, and a few other servers and models break it. Those numbers do not measure what the suite is designed to measure. They are kept and labelled rather than deleted.
Rebuilt from the lab's records by Claude, an AI model. The results above come straight from the automated runs. How these pages are made.
| File | models/qwen3.6-35b-a3b/Qwen3.6-35B-A3B-UD-Q5_K_M.gguf |
| Format | GGUF UD-Q5_K_M (Unsloth imatrix), 27,087,812,896 B |
| Where from | downloaded 2026-08-11 from unsloth/Qwen3.6-35B-A3B-MTP-GGUF, sha256 verified |
| Status | on disk; router preset [qwen3.6-35b] (on demand, not a fleetview seat); benched |
| Hugging Face | https://huggingface.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF |
Where it came from
- Repo / files:
unsloth/Qwen3.6-35B-A3B-MTP-GGUF,Qwen3.6-35B-A3B-UD-Q5_K_M.ggufandmmproj-F16.gguf(models/qwen3.6-35b-a3b/finish.sh,supervise.sh, read 2026-10-23). - Commands:
supervise.sh—curl -sL -C - --connect-timeout 20 --speed-limit 200000 --speed-time 20 -o "$f" "$B/$f"with up to 500 retries; the main file finished through a 6-connection range downloader (pget.log: "done: 27087MB / 27087MB");finish.shfetched the mmproj with 4 connections and checked both against the HF tree API LFS oids. - Date: 2026-08-11, complete 15:03 BST (14:03Z) (file mtime;
.dl_complete15:05 BST). - Size: 27,087,812,896 B (
finish.log; stat 2026-10-23). mmproj 899,283,584 B. - sha256:
8ca8224c5519b50f286b9739f9ab1c2bc0759e02cfb6cbf48a09f0759eb2d319— "OK" against the HF LFS oid (expected.txt,finish.log). mmproj71f3cbc1f7cc0f30d09d41cfa924c0060827ebc33bf15ace7e86661e856f0160OK. - Revision: not on record (resolved
main). - Quant recipe (GGUF header):
general.quantized_by = Unsloth, file_type 17 (Q5_K_M, "UD" dynamic), imatriximatrix_unsloth.ggufoverunsloth_calibration_Qwen3.6-35B-A3B.txt, 510 entries, 77 chunks;nextn_predict_layers = 1(the MTP head; 4 nextn tensors). - Smoke test:
llama-clion buildb10369-6e62ba538, 2026-08-11 15:18 BST, "Prompt: 120.4 t/s | Generation: 42.3 t/s" (smoke.log). - Who / why: chosen over the non-MTP build for the embedded head (a lab note). Alan's request is not on record.
How it was served
| label | dates (UTC) | port / binary | flags that mattered | harness profile | why this label exists |
|---|---|---|---|---|---|
| (none — interactive) | 2026-08-11 → ~2026-08-27 | router, llama.cpp/build/bin/llama-server (b10369 until the 2026-08-25 rebuild to b10618) | preset [qwen3.6-35b]: mmproj, spec-type = draft-mtp, c = 131072; globals n-gpu-layers = 99, jinja = true; launch --models-preset fleet-models.ini --models-max 1 --context-shift | — | daily driver / DSH model (a lab note) |
qwen3.6-35b | 2026-08-27 07:00Z–07:04Z | router, b10618 (eb25b7263) | same preset; -np not set (server default, not on record); no --reasoning-format | extraction (thinking off, max_tokens 2048) | first locked-suite runs: grounding, screening, nullschema |
qwen3.6-35b-tag | 2026-08-31 ~17:00Z–17:30Z | standalone server, b10618 | --ctx-size 16384 --n-gpu-layers 99 --jinja; no MTP, no mmproj | extraction | tagger head-to-head with gemma4-12b and Qwen3-Next under identical standalone flags |
qwen3.6-35b-ag | 2026-09-07 ~18:13Z–18:20Z | standalone server, b10618 | --ctx-size 32768 --n-gpu-layers 99 --jinja; no MTP/mmproj; tool-call gate at 256 tokens | extraction | agentic suites (tools, toolscale, chainwalk) with Qwen3-Next and AgentWorld |
Run-log runtime for all three labels: llama.cpp b10618 (eb25b7263) (runtime_source "evidence"), matching the router and standalone binary. The qwen3.6-35b label also carries one toolcall row that is not this file (Qwen3.6-35B-A3B MLX 4-bit (mlx-community)).
Seats held: default router model and DSH driver from 2026-08-11 (fleet-models.ini "DEFAULT — best balance"); the load-on-startup setting removed 2026-08-20; still a router preset on 2026-10-23 (fleet-models.ini). Never a fleetview seat.
Decisions
- Downloaded and made default 2026-08-11 (family decision log). Kept on disk through every sweep; it has run-log results and remained a router preset.
- Not converted to Q4_0 despite the 2026-09-11 audit flag (no decision on record).
- Not published to HF (a third-party quant).
What went wrong
- Never benched for 16 days. In the router but absent from the run log until 2026-08-27, because it entered as an interactive model.
- Partial row recorded as a result.
scan(record=True)ran mid-suite and logged grounding 86.3 on 73 cases; the complete 82.4 (103/125) was appended after. The run log still shows both, the first unmarked. - Smoke-gate false alarm. The one-tool-call gate with
max_tokens=16reported empty content for this model during the tagger run because reasoning ate the budget. - Harness thinking defaults (2026-08-22). Three board suites failed because the template thinks by default and the harness never sent
enable_thinking: false. - "Exactly one copy" misread on 2026-09-12: the assistant said this 27 GB file was the only copy, missing the 19 GB MLX build in the HF cache.
What happened to it
Downloaded 2026-08-11 → default/DSH driver 2026-08-11 → benched 2026-08-27, 08-31, 09-07 → displaced as interactive model by the MLX build 2026-09-08 → router preset only → on disk 2026-10-23 (27,087,812,896 B). No RESTORE-deleted-*.sh entry needed. Re-downloadable from unsloth/Qwen3.6-35B-A3B-MTP-GGUF (verify against the sha256 above).