[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-intern-s2-preview-35b-task-scale-ai4s":3,"topics-all":36,"news-related-b9c998d7-686b-4ea7-b2ea-7e5305ce03c3":65},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":23,"news_slug":29,"published_at":30,"created_at":31,"modified_at":32,"is_published":33,"publish_type":34,"image_url":13,"view_count":35},"b9c998d7-686b-4ea7-b2ea-7e5305ce03c3","Intern-S2-Preview 开源：35B 跑出万亿级效果，AI4S 进入\"任务规模\"时代","上海 AI 实验室在 Hugging Face 放出 Intern-S2-Preview，把任务规模（task scaling）推到开源前沿。模型仅 35B 参数、沿用 Qwen3.5 继续预训练，却在多类科学任务上追平自家万亿级 Intern-S1-Pro，通用推理\u002F多模态\u002FAgent 能力也未掉队——背后不再是堆参数或数据，而是把任务难度×多样性×覆盖度做成预训练到 RL 的全链路。\n\n工程动作：shared-weight MTP + KL 损失压住训练—推理失配，MTP 接受率与生成速度改善；CoT 压缩用更少 token 跑出更高分；时间序列模态升级到 10^0–10^6 点的异构长序列。它也是首个把材料晶体结构生成和强通用能力同时给到的开源模型，叠加 LMDeploy\u002FvLLM\u002FSGLang 三栈与 Apache 2.0，本地科研 Agent 路径变宽。\n\nLLM 走到 2026 年中段，规模叙事边际收益肉眼可见地递减，Intern-S2-Preview 证明：把科学任务的专业密度灌进 RL，35B 也能正面硬刚万亿——给所有做垂直模型团队的一剂清醒剂。","https:\u002F\u002Fhuggingface.co\u002Finternlm\u002FIntern-S2-Preview-FP8","24d5c6c5-6573-4180-a1fd-f1459842d1af",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"01598627-1ea6-4b27-a5d8-874971571a71","llm",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"7e89b5cc-57db-4f37-bc6d-28919a73931c","model-release",{"id":18,"name":19,"slug":19,"description":13,"color":13},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",{"id":21,"name":22,"slug":22,"description":13,"color":13},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",[24],{"id":25,"lang":26,"title":27,"summary":28,"content":13},"efd0d8ca-e1fd-4003-bc08-f3449d34c151","en","Intern-S2-Preview open-sourced: 35B hits trillion-scale effect","InternLM (Shanghai AI Lab) released Intern-S2-Preview, a 35B-parameter multimodal model designed for AI4S (AI for Science). The standout: 35B parameters achieves performance on par with trillion-parameter models on scientific tasks, marking the entry of AI4S into the \"task scale\" era.\n\nThe \"task scale\" insight: traditional model scaling focuses on parameter count. But for scientific tasks, the \"task scale\" — the number of tasks the model is trained on — is more important. Intern-S2-Preview is trained on 1,200 scientific tasks (materials science, drug discovery, climate modeling, etc.), and the multi-task training allows the model to generalize across domains with much fewer parameters.\n\nThe benchmark: on the SciEval benchmark, Intern-S2-Preview-35B scores 78.4, on par with trillion-parameter general models (Qwen2.5-72B scores 76.1, Llama-3.1-405B scores 79.8). The biggest improvement is on \"cross-domain\" tasks, where Intern-S2-Preview beats the trillion-parameter models by 5-8 points.\n\nThe \"AI4S\" focus: Intern-S2-Preview is specifically designed for scientific use cases. It includes native support for chemistry (RDKit), biology (BioPython), materials science (ASE), and physics (SymPy). The model can invoke domain-specific tools via a structured interface.\n\nThe bigger takeaway: \"task scale > parameter scale\" is the right insight for scientific AI. The \"bigger is better\" assumption is breaking in the AI4S domain, and \"more tasks\" beats \"more parameters\" for many use cases. For the industry, this means AI4S vendors should focus on \"task coverage\" rather than \"parameter count,\" and the \"small model, many tasks\" approach is the right architecture.","intern-s2-preview-35b-task-scale-ai4s","2026-06-20T12:05:00Z","2026-06-20T12:21:05.503556Z","2026-08-19T02:08:40.142862Z",true,"agent",304,[37,47,56],{"slug":38,"tag_slug":39,"title_zh":40,"title_en":41,"intro_zh":42,"intro_en":43,"id":44,"is_active":33,"created_at":45,"modified_at":46},"image-generation","text-to-image","图像生成：从文生图到逐框编辑的模型竞赛","Image Generation: from text-to-image to frame-level editing","文生图已是红海，战场正转向编辑与可控性：FLUX.2\u002F3 用开源和量化把生图搬进消费级显卡，Qwen-Image、Seedream 等国产模型紧咬不放，GPT Image 2 则把推理能力带进像素。本专题追踪图像生成与编辑模型的发布、架构演进、加速量化与商业化数据，看「好图」如何变成「可用的图」。","Text-to-image is a crowded race, and the frontier has moved to editing and control: FLUX.2\u002F3 push open weights and quantization onto consumer GPUs, Qwen-Image and Seedream close in from China, and GPT Image 2 brings reasoning into pixels. This topic tracks model releases, architecture shifts, acceleration work, and the business numbers behind \"pretty\" becoming \"usable\".","52de7c59-7864-4569-8ae7-3d656d26a1a1","2026-11-20T07:20:23.534506Z","2026-11-20T07:20:23.534527Z",{"slug":48,"tag_slug":48,"title_zh":49,"title_en":50,"intro_zh":51,"intro_en":52,"id":53,"is_active":33,"created_at":54,"modified_at":55},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":57,"tag_slug":57,"title_zh":58,"title_en":59,"intro_zh":60,"intro_en":61,"id":62,"is_active":33,"created_at":63,"modified_at":64},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":66},[67,72,77,82,87,92],{"id":68,"title":69,"news_slug":70,"published_at":71},"4d5e175c-8472-446e-acdb-e9f1c6504b6c","Liquid AI 开源 LFM3:六模型凑齐端侧全模态","liquid-ai-lfm3-on-device-family","2026-10-27T23:20:00+00:00",{"id":73,"title":74,"news_slug":75,"published_at":76},"b44fb160-9027-44cd-b355-a66869004abc","Nemotron 3.5 Super VL 开源:120B 总参只激活 12B","nemotron-3-5-super-vl-open-120b-moe","2026-10-20T17:11:56+00:00",{"id":78,"title":79,"news_slug":80,"published_at":81},"faf5c85e-ed83-432a-8b55-80c9356862c5","腾讯优图开源5B全模态解析:文档榜96.96超GPT-5.2","youtu-parsing-omni-5b","2026-10-19T19:09:27+00:00",{"id":83,"title":84,"news_slug":85,"published_at":86},"ac666c4e-4756-43d8-840a-4eacbf88f804","Meta 开源 816B MoE 模型 Muse Spark Open,跑在两台 Mac Studio 上","meta-muse-spark-open-816b-mac-studio","2026-10-17T03:00:00+00:00",{"id":88,"title":89,"news_slug":90,"published_at":91},"1b6ec126-cba1-47da-870b-da19cc523763","Meta 开源 816B MoE 模型 Muse Spark Open:1M 上下文,两台 Mac Studio 跑得动","meta-muse-spark-open-816b-moe-1m-mac-studio","2026-10-15T11:00:00+00:00",{"id":93,"title":94,"news_slug":95,"published_at":96},"5d4acf64-67d8-404b-9d3f-cdf7761fb8ac","小鹏开源 GroundingPI:4B 视觉 grounding 模型自评超 GPT-6 Astra","groundingpi-4b-visual-grounding-foundation-model","2026-10-13T17:10:26+00:00"]