HuatuoGPT-3-27B
Medical LLM built on Qwen3.8-27B using OnePO - single-stage policy optimization that adapts a base model to medicine with RL only, no domain SFT.
Why it matters
Skips the expensive domain SFT stage entirely: teacher responses give temporary guidance then retire. Releases training code, the OnePO-Medical-20K RL dataset, and an 8B rubric grader - a full recipe for SFT-free domain adaptation.
What you could build with it
Self-host HuatuoGPT-3-27B as a private clinical scribe for Indian clinics: transcribe consults, draft discharge summaries, then use the released OnePO pipeline to adapt it to local-language records without a big SFT budget.
Does it hold up?
Self-hosting is feasible since weights, OnePO training code, medical RL data, and the grader are all public; there are no independent deployments, fine-tunes, or benchmark reproductions yet.
Built with HuatuoGPT-3-27B
- Open-source healthcare AI roundupwriteup · Healthcare software roundup citing the HuatuoGPT-3 release and its open weights, OnePO code, medical RL data, and grader.
- Lemmy release discussiondiscussion · Community thread discussing the HuatuoGPT-3 release and what the open components enable for builders.
- Model release feed entrywriteup · Release feed tracking the HuatuoGPT-3 model drop for the community.
Learn more
First spotted on github: source.
More AI models
ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71DeepSeek V4-Flash official API: public beta with upgraded agent capabilities, Responses API and Codex supportDeepSeek's official V4-Flash is now a public beta API with massively upgraded agent capabilities - benchmark scores…models · JEV 0.69Mistral Forge: enterprise platform for training and continuously improving proprietary modelsMistral AI launched Forge (Sep 28) — a full-lifecycle model training platform (pre-training, SFT, DPO/ODPO, RL…models · JEV 0.69kev: open, trainable Jev-like family of small decision models on Qwen3.5/3.8A family of small decision models (0.8B to 27B) built on Qwen3.5/3.8 that reproduces Jev's prefill-only typed-decision…models · JEV 0.68GPT-Synopsys: OpenAI and Synopsys multi-year deal to build an AI model that operates EDA toolsOpenAI and Synopsys signed a multi-year agreement on Sep 30 to jointly develop GPT-Synopsys, a specialized model that…models · JEV 0.66
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.