8B-parameter clinical multimodal model EVLF-FM beats clinicians, published in Nature
YogurtExternal7923 · reddit · 2026-10-31
EVLF-FM, an 8B-parameter clinical multimodal vision-language foundation model published in Nature, surpasses clinicians in multimodal understanding across radiology, dermatology and more. Architecture: three encoders (general vision, disease-level, pixel-level) feeding an LLM decoder; two-stage training with SFT (LoRA) then GRPO reinforcement learning with clinical-reference rewards; a rigorous CoT data curation pipeline (generation, uncertainty filtering, quality scoring, refinement). Covers 330K pathology, 323K XR, 284K MRI and other large-scale training data; applications include diagnosis, medical grounding and VQA — frontier-level medical AI without a data center.
More from Models
- Codex 28-day challenge teases day-28 reset: OpenAI may be saving a new model for the finale — haider1 · 2026-10-31
- GPT-6 ASTRA LAW spotted in OpenAI's API model list, launch looks imminent — imjustnewatai · 2026-10-31
- Grok now accepts Alipay and UnionPay, opening paid access for Chinese users — op7418 · 2026-10-31
- Asking DeepSeek and Hy to predict their own next models yields two very different roadmaps — lucky-adventurer- · 2026-10-31
- Tencent Hunyuan 4 Flash rumored; Reddit user says cheap Flash models shine for agent workflows — Positive_Hall8510 · 2026-10-31
- Rumor: A Single 7GB Shard of 'Fable 5.5' Weights Leaked Out of 8TB Total, Likely Fake — power97992 · 2026-10-31