Winnow-12B
72.2#54 of 261 on the open-weights board · Capability #68
Model guide · Gemma 4 12B decision fine-tune · GGUF · decisions, chat and vision · updated October 26, 2026
One local model that decides, chats and looks at a screenshot.
Winnow-12B fine-tunes Google’s Gemma 4 12B for Jev-style typed decisions and serves them from a llama.cpp-based server that also answers ordinary chat and takes images through an optional projector. Its author tested the 8-bit build with a 64K context and vision on a 16 GB RTX 5070 Ti. JevBench v1.7.48 ranks it #54 of 261 open-weights systems, just above Jev’s 71.4 reference, and the strongest system on the board that also chats.
Winnow-12B is a third-party project by EldanRing. Jev AI does not host it; this page is a guide to what it is and how to run it yourself.
Try it on this page
Edit the example or paste your own text, then run it. Sign in to run: new accounts get 20 welcome credits, and a failed run costs nothing.
Jev AI does not host Winnow-12B, so this editor runs Jev 1.14 on the same kind of decision. Compare the answers with your own Winnow-12B setup.
Support ticket triage · Jev 1.14
1 Input
Hi, this is the third time I am writing. Our payouts have failed every day since Monday and my team cannot pay suppliers. I already re-entered the bank details twice. If this is not fixed today we will have to move to another provider.
2 Questions
One benchmark measured every system under one method. It ranks 261 open-weights systems on one board and 54 hosted APIs on another; Jev, a hosted API, is ranked on the second and shown on the first as an unranked reference. The full board is on the JevBench results page.
Winnow-12B
72.2#54 of 261 on the open-weights board · Capability #68
Jev 1.13.0
71.4Open-weights reference row · #18 of 54 on the API board · Capability #15
IntelligenceHow often it picks the right answer, half from sealed decisions
CalibrationWhether 0.8 really means about 80%
SpeedMeasured response time
Open half 59.8 · sealed half 59.2; chance-corrected by request type: choice 70.1, yes/no 51.5, score 57.0. The composite also weighs a fourth axis, cost, which these bars leave out. Self-hosted latency is adjusted by the benchmark to approximate production load.
Winnow-12B is a LoRA fine-tune of google/gemma-4-12B-it merged into the weights and released as two GGUF files: an 8-bit Q8_0 build recommended for 16 GB GPUs and a 16-bit BF16 build. It runs in the author’s llama.cpp-based inference server; no conversion step or separate adapter is needed.
For decisions the server fills in the shared state once, forks one branch per question and reads the logits of the answer tokens without generating text, the same idea as Jev’s typed questions. It supports noul, choice and score. The same loaded model serves chat on /v1/chat/completions, and the optional projector adds image input. It states that it is not affiliated with or endorsed by TypeSafe or Google.
The training data is private. The model card describes curated typed-decision examples covering routing, policy and rule application, evidence selection, workflow decisions, ordinal judgments, entailment, paraphrase and answerability, followed by a refinement pass.
JevBench v1.7.48 ranks Winnow-12B Q8 #54 of 261 open-weights systems at 72.2, measured on an RTX 6000 at a 0.12s raw median per decision. Intelligence is 59.5, with 59.8 on the open half and 59.2 on the sealed half; calibration is 83.0 and speed 86.2. Jev, the board’s unranked reference, scores 71.4 with intelligence 63.6 and calibration 90.6. Winnow was second, above Jev, in the previous release on the old pool; on the current board it edges Jev’s reference score on the composite thanks to its cost score since Benchmark Heaven’s 29 October GPU-time cost basis, while Jev stays well above it on the Capability ranking, where Winnow is #53.
Its model card also reports a second suite, Kev-v9 clean, with 1,046 items, where the author scores Jev at 87.00% and Winnow Q8 at 81.55%, and a tie at 85.71% on the 231-item public JevBench subset. Those are the author’s measurements, not JevBench’s.
Download the Q8_0 GGUF and, if you want image input, the projector; then start the author’s inference server. One process serves decisions and chat.
# Weights: huggingface.co/EldanRing/Winnow-12B (Q8_0 ~12.7 GB, BF16 ~23.8 GB, projector 175 MB)
git clone https://github.com/EldanRing/winnow-inference
cd winnow-inference
# Follow README.md to build the llama.cpp-based server and point it at the GGUF.
curl http://localhost:8080/v1/systemone \
-H "Content-Type: application/json" \
-d '{"state": "Refund request: item returned unopened 40 days after delivery.",
"questions": {"eligible": {"type": "noul", "instructions": "Is this within a 30-day return window?"}}}'From the project’s own documentation and JevBench v1.7.48.
Winnow does more on one GPU than anything else on this site, so the useful comparison is on your own cases and your own constraints. Run them on Jev here with no setup and keep the answers; if Winnow agrees on the decisions you care about and you can run it, a local model that also chats and reads images may still be the better fit.
An Apache-2.0 fine-tune of Google’s Gemma 4 12B by EldanRing that answers Jev-style noul, choice and score questions from a llama.cpp-based server, and also serves chat and image input from the same loaded model.
The author tested the 8-bit Q8_0 build, about 12.7 GB, with full offload on a 16 GB RTX 5070 Ti at a 64K context. The BF16 build, about 23.8 GB, needs more memory or offloading.
Not on the answers in JevBench v1.7.48: it is #54 of 261 open-weights systems at 72.2, a little above the 71.4 of Jev, shown there as an unranked reference, thanks to its cost score, but with intelligence 59.5 against 63.6. It was ahead of Jev in v1.5.6, on the old pool.
Yes, with the optional 175 MB vision projector loaded alongside the GGUF. Jev is text only.
Its server exposes /v1/systemone for typed decisions with noul, choice and score questions, and /v1/chat/completions for chat. Test your own requests against it before switching an integration.
Every figure on this page is quoted from a published source and was read on October 26, 2026.
Winnow-12B and every other product named on this page belongs to its respective owner and is not affiliated with Jev AI. Figures are quoted from the sources above as they read on October 26, 2026 and can change.