StepFun· @StepFun_ai · X·· 2026-05-29AI 评分72
AI 导读
阶跃星辰发布的多模态模型 Step 3.7 Flash 已在 ModelScope 平台上线。该模型采用 198B 总参数的 MoE 架构,每个 token 激活 11B 参数,推理速度最高达 400 tok/s,支持 256K 上下文窗口,并提供低、中、高三个推理级别以平衡速度与效果。其在 ClawEval-1.1 榜单位列第一(67.1分),在 SWE-bench Pro 上排名第二(56.3分)。模型具备原生多模态能力,由语言骨干与视觉编码器组成,原生支持解析密集UI、图表及财报。该模型采用 Apache 2.0 协议开源,并兼容 vLLM 等多种推理框架。
正文 · 原文
Excited to see Step 3.7 Flash available on @ModelScope2022 🚀
Can’t wait to see what builders create with it!
Thrilled to welcome Step 3.7 Flash landing on ModelScope, a 198B sparse MoE VLM from @StepFun_ai 🔥🤖 https://modelscope.cn/collections/stepfun-ai/Step-37-Flash 198B total, 11B active per token, up to 400 tok/s. 256K context. Three reasoning levels (low/medium/high) to balance speed, cost, and depth. 🥇 #1 on ClawEval-1.1 (67.1), leading Kimi K2.6 by 9.3pts 🥈 #2 on SWE-bench Pro (56.3) 🥇 #1 on SimpleVQA Search (79.2) 🖼️ Multimodal: 196B language backbone + 1.8B vision encoder. Parses dense UI, charts, and financial reports natively. 📄 Apache 2.0. vLLM / SGLang / llama.cpp / Transformers supported. Hermes Agent / OpenClaw / Kilo Code compatible. Runs locally on Mac Studio or DGX Station with 128GB+ memory.在 X 查看被引用的帖子
来源:StepFun · x.com