[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-sage-a1-5-agentic-scientific-discovery":3,"topics-all":39,"news-related-7d735b5c-c7a8-4a8d-afac-abd3034170e7":57},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":25,"news_slug":32,"published_at":33,"created_at":34,"modified_at":35,"is_published":36,"publish_type":37,"image_url":14,"view_count":38},"7d735b5c-c7a8-4a8d-afac-abd3034170e7","Sage-A1.5开源:35B智能体GAIA 95.1","上海AI实验室InternScience发布35B MoE智能体模型Sage-A1.5,基于Qwen3.6-35B-A3B后训练,新增并行子智能体编排、思维链内自查验证与开放式科学发现三类能力。官方评测GAIA 95.1,称在其对比集中超过所有更大模型;权重以Apache 2.0开源,另放出FP8与GGUF量化版。","10 月 20 日,上海AI实验室的 InternScience 团队把 Sage-A1.5 的技术报告挂上 arXiv(编号 2610.28578),配套模型卡与全套权重随后在 Hugging Face 更新到位。这是一个 35B 的 MoE 智能体模型,底座为 Qwen3.6-35B-A3B,官方定位是 A1 的升级:不只解决单个任务,而是学会「组织工作、并在执行中修正自己」。论文上线一天,HF 论文页攒下 58 个赞。\n\n## 能力组合:三个新引擎\n\n报告把升级拆成三块。第一,智能体编排:给模型暴露一个 spawn_subagent 工具,它就能把任务拆给多个并行子智能体,每个子智能体有独立上下文和工具权限,互不干扰地各查各的,最后汇总结论。第二,验证模式:通过 chat_template_kwargs 打开开关后,模型会在自己的思维链中途暂停,复核哪些结论站得住、哪些路径走死了,再在同一回合里换方向——不需要外挂一个批评者模型。第三,开放式科学发现:从回答预设问题走向探索式研究,产出可检查的研究工件,官方用多学科科研任务做了展示。\n\n## 官方跑分:35B 掂量更大的模型\n\n模型卡的评测数字全部来自官方自报,社区尚无第三方复现,先按官方口径看:GAIA 拿到 95.1,官方称高于其对比集里所有更大规模的模型,对比对象包括 Kimi-K3、GLM-5.3 与 GPT-5.6-sol;XBench-DS-2510 92.0、IFBench 81.6、FrontierScience-Olympiad 81.0、FrontierScience-Research 48.3、OmniDocBench 91.5。同级 35B 档排第一的项目包括 SWE-bench Pro 61.2、BrowseComp 81.7、带工具的 HLE 55.1。相对上一代 A1,工程类提升显著:SWE-bench Pro +17.6、SWE-bench Multilingual +14.0、MLE-Lite +21.3、PinchBench +13.1。\n\n## 部署侧:一张 GPU 与 GGUF\n\n推理配置写得务实:vLLM 单卡即可起 262K 上下文服务,支持 MTP 投机解码(SGLang 走 EAGLE,vLLM 走 mtp 配置);纯文本场景可跳过视觉编码器,把 KV cache 省下来——模型卡标签里有 vlm,这是个带视觉的模型。本地玩家也有份:FP8、Q4_K_M、Q8_0、F16 四种量化版本全部放出,许可证 Apache 2.0。\n\nSage 的看点不在单个跑分,而在路线:旗舰模型靠规模碾压时,35B 这条线在赌「会编排、会自我验证」比「更大」更划算。如果官方数字经得起第三方复现,开源科研智能体的门槛会再降一档。手头有闲置显卡的话,这套权重值得烧一晚上。\n\n[技术报告](https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.28578) · [模型卡](https:\u002F\u002Fhuggingface.co\u002FInternScience\u002FSage-A1.5)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.28578","7437aeb9-930c-4866-a2e9-48003c1a792b",[11,15,19,22],{"id":12,"name":13,"slug":13,"description":14,"color":14},"6ad31a14-c0da-42df-81fd-564281f768db","agentic-ai",null,{"id":16,"name":17,"slug":17,"description":18,"color":14},"9112951a-2abb-4214-b63a-385ec7afb2ba","ai-for-science","AI for Science 专题：追踪 AI 在生命科学、化学材料、物理世界模型等科学方向的关键突破",{"id":20,"name":21,"slug":21,"description":14,"color":14},"7e89b5cc-57db-4f37-bc6d-28919a73931c","model-release",{"id":23,"name":24,"slug":24,"description":14,"color":14},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",[26],{"id":27,"lang":28,"title":29,"summary":30,"content":31},"cd8e97ac-008c-4599-9c4f-513253dfd8ad","en","Sage-A1.5: a 35B agentic model scoring 95.1 on GAIA","Shanghai AI Lab open-sources Sage-A1.5, a 35B MoE agentic model with parallel subagent orchestration and in-CoT verification; GAIA 95.1, Apache 2.0.","On October 20, the InternScience team at Shanghai AI Laboratory posted the technical report for Sage-A1.5 on arXiv (2610.28578), with the model card and the full weight family landing on Hugging Face shortly after. It is a 35B Mixture-of-Experts agentic model post-trained from Qwen3.6-35B-A3B, and the team frames it as an upgrade over A1: the model no longer just solves individual tasks, it learns to organize work and revise its own approach mid-execution. Within a day, the Hugging Face paper page had collected 58 upvotes.\n\n## Capability composition: three new engines\n\nThe report splits the upgrade into three parts. Agentic orchestration: expose a spawn_subagent tool and the model fans tasks out to parallel subagents, each with its own context window and tool access, working without mutual interference before synthesizing findings. Verification mode: enabled through chat_template_kwargs, the model pauses inside its own chain of thought to check which conclusions hold and which search paths are dead, then switches direction in the same turn — no external critic model required. Open scientific discovery: the model moves from answering predefined questions to exploratory research, connecting scientific inquiry with interactive analysis and inspectable artifacts, demonstrated across multidisciplinary scientific applications.\n\n## Official benchmarks: 35B weighing in against bigger models\n\nEvery number in the model card is vendor-reported; the community has no third-party reproduction yet, so read them as official figures. Sage-A1.5 reaches 95.1 on GAIA, which the team says beats every larger-scale model in its comparison set, including Kimi-K3, GLM-5.3 and GPT-5.6-sol. It posts XBench-DS-2510 at 92.0, IFBench at 81.6, FrontierScience-Olympiad at 81.0, FrontierScience-Research at 48.3 and OmniDocBench at 91.5. Among comparable 35B-class models it ranks first on SWE-bench Pro (61.2), BrowseComp (81.7) and HLE with tools (55.1). Against the previous A1 generation, engineering workloads improved the most: SWE-bench Pro +17.6, SWE-bench Multilingual +14.0, MLE-Lite +21.3, PinchBench +13.1.\n\n## Deployment: one GPU and GGUF\n\nThe serving story is pragmatic. A single GPU runs a 262K-context vLLM deployment, with multi-token-prediction speculative decoding available (EAGLE under SGLang, mtp config under vLLM). Text-only workloads can skip the vision encoder to free KV cache — the model card tags it as a VLM. Local users get FP8, Q4_K_M, Q8_0 and F16 GGUF editions, all under Apache 2.0.\n\nThe interesting part is not any single score but the bet: while frontier models grind forward on scale, this 35B line wagers that orchestration and self-verification buy more than raw size. If the official numbers survive independent reproduction, the entry bar for open scientific agents drops another notch.\n\n[Technical report](https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.28578) · [Model card](https:\u002F\u002Fhuggingface.co\u002FInternScience\u002FSage-A1.5)","sage-a1-5-agentic-scientific-discovery","2026-10-24T23:10:02Z","2026-10-24T23:10:05.792765Z","2026-10-24T23:10:05.792777Z",true,"agent",32,[40,48],{"slug":17,"tag_slug":17,"title_zh":41,"title_en":42,"intro_zh":43,"intro_en":44,"id":45,"is_active":36,"created_at":46,"modified_at":47},"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":49,"tag_slug":49,"title_zh":50,"title_en":51,"intro_zh":52,"intro_en":53,"id":54,"is_active":36,"created_at":55,"modified_at":56},"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":58},[59,64,69,74,79,84],{"id":60,"title":61,"news_slug":62,"published_at":63},"48726cf0-6f0c-4b09-841e-06aa8b0ffcfd","上海AI Lab Intern-Decision:多模态决策模型,4B 微超 Jev 90.02%","intern-decision-qwen3-5-multimodal-decision-model","2026-10-18T11:13:06+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"4cc278cf-6b53-4722-81ac-d1909b50e6a9","Meta 开源 816B 模型，Mac 可本地跑","meta-muse-spark-open-816b-mac","2026-10-15T23:10:00+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"08872b65-a25f-41f2-877c-a6f265e3c72f","LFM2.5-24B开源:2B激活跑赢Qwen3.6-35B","liquid-ai-lfm2-5-24b-a2b","2026-10-14T23:10:00+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"63ae3eda-4994-4f9f-a812-489d9831a9ca","Clef-omni 开源:决策模型学会听音看片","clef-omni-multimodal-decision-model","2026-10-12T15:20:00+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"ca3472cb-0638-440d-ae0e-f6c3eac95cf7","亚马逊开源2B决策模型:权重、数据、训练脚本全放","amazon-strands-decider-2b","2026-10-04T21:15:00+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"52f192a9-5630-435d-a76a-237e2c94eda5","AstaBrief 开源:8B 写科研报告快 3.5 倍","astabrief-8b-scientific-report-model","2026-10-04T15:06:06+00:00"]