跳到正文
Aran Komatsuzaki· @arankomatsuzaki · X·· 2 小时前AI 评分58
AI 导读

Ant Ling 发布开源原生多模态医疗大模型 LingSante-3.1-air-VL,总参数 124B、激活参数 5.5B,结合医学图像理解、报告生成与医疗推理,并支持多模态搜索和工具调用。引用内容称其在 AFUMED-Derm、AFUMED-Report、MIMIC-CXR 和医疗 VQA 四项评测中领先 GPT-6.1 Sol 与 Gemini 3.8 Flash,在 BrowseComp-VL 和 AndroidWorld 上也高于这两款模型。模型已在 Hugging Face 开放。

正文 · 原文

Ant Ling presents LingSante-3.1-air-VL, an opensource natively multimodal LLM with 5.5B/124B params

  • SotA on medical / healthcare benchmarks among flash-tier models
  • Competitive performance on various non-medical tasks (e.g. OmniDocBench, BrowseComp) among flash-tier models

https://huggingface.co/inclusionAI/LingSante-3.1-air-VL

引用Ant Ling@AntLingAGI
Meet LingSante-3.1-air-VL, an open-source multimodal medical model with 124B total and 5.5B active parameters. It combines medical image understanding, report generation and medical reasoning with multimodal search and tool use. In our evaluation, it leads the compared models on AFUMED-Derm, AFUMED-Report, MIMIC-CXR and medical VQA (averaged across benchmarks), ahead of GPT-6.1 Sol, Gemini 3.8 Flash on all four. It also demonstrates strong general-purpose capabilities, scoring higher than GPT-6.1 Sol on BrowseComp-VL and Gemini 3.8 Flash on AndroidWorld, alongside competitive performance on MMSearch. Explore LingSante-3.1-air-VL and build with it.
在 X 查看被引用的帖子

来源:Aran Komatsuzaki · x.com