Qwen3.5-4B fine-tuned in 27.5 min on one GPU lifts MedQA accuracy to 74.5%
MaziyarPanahi · x · 2026-11-13
Using Unsloth's decision-model recipe, Qwen3.5-4B was fine-tuned in 27.5 minutes on a single GPU. On the unseen MedQA test set (USMLE-style, 4 options), overall accuracy rose from 66.7% to 74.5%, and calibration error dropped from 0.088 to 0.024—a strikingly cheap training run with sizable gains.
More from Models
- Ditto rolls out uncensored inference to all users — bittingthembits · 2026-11-13
- Google says Gemini 4 Hydrogen expands query fan-out and the range of sites Search links to — OnlineInference · 2026-11-13
- 124B model in 22GB: Cactus Quants hits 1.418 bits/weight, 47 tok/s on an M5 Max — pbaylies · 2026-11-13
- PostTrainBench: Fable 5.2 ties 5.1 at 44.6%, speculating Anthropic capped AI research ability — scaling01 · 2026-11-13
- Gemini 4 Hydrogen lands in Devin, scoring 38.2% on FrontierCode 2.0 behind GPT-6.1 Sol — scaling01 · 2026-11-13
- Gemini 4 Carbon comes within 1.9 points of Fable 5.2 on Code4Scene at ~1/10 the cost — ZhitingHu · 2026-11-13