aifollow.news
返回 Rohan Paul
Rohan Paul· @rohanpaul_ai · X· · 原发布时间 精选AI 评分60

Ant Group 发布医疗模型 LingSante-3.1-flash,开放模型权重

AI 导读

LingSante-3.1-flash 是总参数量为 5600 亿的医疗混合专家模型,每个 token 激活 250 亿参数。发布方称,它在 HealthBench Professional 等三项评测中超过 GPT-6 Astra、Claude Opus 5.5 和 Gemini 4 Argon;模型权重现已开放。

推荐理由

医疗模型权重已开放,且发布方给出了与其他模型比较的具体评测结果;这些成绩仍属于发布方的主张。

原帖引用 · AI 翻译

@rohanpaul_ai
译文 · 以原文为准

https://t.co/QSpKY4Da5G

https://t.co/otJbpZSAYg

回复Rohan Paul@rohanpaul_ai
译文
蚂蚁集团发布了 LingSante-3.1-flash,这是一款拥有 5600 亿参数的医疗领域 MoE 模型;它在 HealthBench Professional、HLE、BrowseComp 和 EHR-Complex 上的表现超过 GPT-6 Astra、Claude Opus 5.5 和 Gemini 4 Argon。 HealthBench Professional 涵盖对真实临床医生向聊天机器人提出的实际请求的回答。EHR-Complex 测试智能体查询医院病历数据库的能力。 模型权重已在 Hugging Face 上发布,采用 MIT 许可证。 蚂蚁从预训练阶段起就混合使用医疗和通用数据,将医疗推理、工具使用、搜索和临床医生式行为作为独立技能进行训练;针对经过医生和验证器核查的任务,每项技能都有各自的强化学习奖励,随后通过多教师同策略蒸馏将这些专门模型合并为一个模型。
原文

Ant Group released LingSante-3.1-flash, a 560B-parameter medical MoE model, and it tops GPT-6 Astra, Claude Opus 5.5 and Gemini 4 Argon on HealthBench Professional, HLE, BrowseComp and EHR-Complex. HealthBench Professional cover answering real clinician requests that real doctors bring to chatbots. EHR-Complex tests an agent querying a hospital records database. weights on Huggingface, with MIT-license. Ant mixed medical and general data from pretraining onward, trained medical reasoning, tool use, search and clinician-style behavior as separate skills with their own RL rewards on physician- and verifier-checked tasks, then merged those specialists into one model through multi-teacher on-policy distillation.

在 X 查看这条帖子 ↗
相关上下文Ant Ling@AntLingAGI
译文
今天,我们发布 LingSante-3.1-flash,这是一款开放权重的 MoE 模型,总参数量为 5600 亿,每个 token 激活 250 亿参数。它在保留通用能力的同时,整合了多种医疗能力,包括复杂医疗推理、长周期工作流程执行,以及专业行为对齐。 LingSante-3.1-flash 超越医疗问答,转向解决现实世界的医疗保健问题。该模型在 HealthBench Professional(65.72)、HLE(仅文本,使用工具:72.22)和 BrowseComp(提供上下文:94.10)上领先,在这 3 项测试中均超过 GPT-6 Astra、Claude Opus 5.5 和 Gemini 4 Argon。LingSante-3.1-flash 在专业的长周期医疗工作流程中也表现出色,例如 PhysicianBench(57)和 EHR-Complex(74.40)。 综合这些结果,LingSante-3.1-flash 成为医疗保健领域领先的开放权重模型。 模型权重现已提供。探索 LingSante,并用它进行开发。 更多内容见下文 ↓
原文

Today, we release LingSante-3.1-flash, an open-weight MoE model with 560B total parameters and 25B activations per token. It unifies diverse medical capabilities while retaining general capabilities, including complex medical reasoning, long-horizon workflow execution, and professional behavior alignment. LingSante-3.1-flash moves beyond medical QA toward real-world healthcare problem. The model leads on HealthBench Professional (65.72), HLE (text-only, with tools: 72.22) and BrowseComp (with context: 94.10), outperforming GPT-6 Astra, Claude Opus 5.5 and Gemini 4 Argon on all three. LingSante-3.1-flash also performs well on professional long-horizon medical workflows, such as PhysicianBench (57) and EHR-Complex (74.40). Together, these results establish LingSante-3.1-flash as the leading open-weight models for healthcare. The weights are available. Explore LingSante and build with it. More below ↓

在 X 查看这条帖子 ↗

来源:Rohan Paul · x.com

模型
发现内容有误?提交纠错