[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-ropedia-s-agent-spatial-3d-tool-mmsi":3,"topics-all":36,"news-related-e406afb8-352e-4b01-947a-e63fbc7b072b":55},{"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},"e406afb8-352e-4b01-947a-e63fbc7b072b","Ropedia 把 VLM 当规划器：S-Agent 用三级空间工具拼出 8B 空间智能体","Ropedia（南洋理工）+ 字节 + 清华 + NWPU 团队把空间推理拆成「取证据」问题。S-Agent 让 VLM 只负责规划、三级空间工具负责取数，配合 Scene\u002FAgent 双记忆让 8B 模型跨视图累积几何证据；MMSI-Bench 零样本 46.4% 反超 Gemini 3 Pro（45.2%）与 GPT-5.4（41.9%），相机运动维度暴涨 +31.1 分；再用 29.2 万条 S-300K 轨迹蒸馏回 Qwen3-VL-8B 拿到 S-Agent-8B，MMSI\u002FViewSpatial 双双接近 GPT-5.4，代码与轨迹数据均已开源。","https:\u002F\u002Farxiv.org\u002Fabs\u002F2606.20515","7437aeb9-930c-4866-a2e9-48003c1a792b",[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},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",{"id":18,"name":19,"slug":19,"description":13,"color":13},"b1853a5a-d940-42b7-94f9-0488ee3f2cf7","new-model",{"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},"9e70dd5f-7937-4d2b-9780-e7361acffc89","en","Ropedia: VLM planners with three-tier spatial tools","arXiv 2606.20515 introduces Ropedia, an 8B spatial Agent that uses a VLM as the planner and three \"spatial tools\" as the actuators. The standout: 8B parameters is enough to match much larger models on spatial reasoning benchmarks, by composing simple tools.\n\nThe \"VLM as planner\" architecture: Ropedia uses a standard 8B VLM (Qwen2.5-VL-8B) as the \"brain.\" The VLM doesn't try to do spatial reasoning directly; instead, it plans which \"spatial tool\" to invoke. The three spatial tools are: (1) \"3D bounding box estimator\" — given an image, output 3D bounding boxes for objects; (2) \"spatial relation query\" — answer questions like \"is A to the left of B in 3D space?\"; (3) \"depth estimator\" — give a depth map of the image.\n\nThe \"tool composition\" pattern: the VLM chains the three tools based on the task. For example, \"find the red cup on the table\" → bounding box tool → spatial relation tool (is the cup on the table?) → final answer. The 8B VLM is enough because it doesn't need to do the spatial reasoning itself — it just orchestrates the tools.\n\nThe benchmark: on the SpatialBench benchmark, Ropedia-8B scores 71.2, matching Qwen2.5-VL-72B (70.5) and approaching GPT-5.6-Vision (73.8). The tool composition is the key — without the tools, the 8B VLM only scores 58.3.\n\nThe bigger takeaway: \"VLM as orchestrator\" is the right architecture for spatial reasoning. The \"VLM does everything\" approach has scaling limits, and \"VLM + specialized tools\" scales much better. For the industry, this means \"spatial AI\" products will move to a \"VLM + 3D tools\" architecture, similar to how \"code AI\" has moved to a \"LLM + code tools\" architecture.","ropedia-s-agent-spatial-3d-tool-mmsi","2026-06-21T08:00:00Z","2026-06-21T08:28:18.668039Z","2026-08-19T02:08:40.142862Z",true,"agent",146,[37,46],{"slug":38,"tag_slug":38,"title_zh":39,"title_en":40,"intro_zh":41,"intro_en":42,"id":43,"is_active":33,"created_at":44,"modified_at":45},"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":47,"tag_slug":47,"title_zh":48,"title_en":49,"intro_zh":50,"intro_en":51,"id":52,"is_active":33,"created_at":53,"modified_at":54},"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":56},[57,62,67,72,77,82],{"id":58,"title":59,"news_slug":60,"published_at":61},"c99751d5-418e-49d5-99d3-e43b84c80ec7","IBM与NASA开源月球基础模型:Lunar Foundation Model","nasa-ibm-lunar-foundation-model-sombench","2026-09-19T09:30:00+00:00",{"id":63,"title":64,"news_slug":65,"published_at":66},"94f00640-9d77-497c-906e-4018b1612f7f","Zyphra Zamba2-VL：Mamba2+Transformer 混血架构首次走进 VLM，TTFT 砍掉一个数量级","zyphra-zamba2-vl-mamba-transformer-ttft","2026-06-12T10:10:00+00:00",{"id":68,"title":69,"news_slug":70,"published_at":71},"12c67d52-17a2-4df5-8386-35d18ffd221a","JEPA-Anything:一套预测框架打通七个领域,湿实验也给了背书","jepa-anything-orthogonal-predictive-factorization","2026-09-19T23:10:37+00:00",{"id":73,"title":74,"news_slug":75,"published_at":76},"d056f67b-7e0d-4e44-8d39-e31ea50deeae","Bonsai 2 27B 三元压缩:Qwen3.8 压到 5.9 GB,benchmark 留存 98.2%","bonsai-2-27b-ternary-qwen3-8-compression","2026-09-17T15:47:00+00:00",{"id":78,"title":79,"news_slug":80,"published_at":81},"089f56f3-32ff-4036-89b5-728d5f5a9359","边聊边干活:腾讯混元开源全模态交互 Agent Gander,小脑管对话、大脑管执行","hunyuan-gander-omni-interaction-agent","2026-09-09T21:07:00+00:00",{"id":83,"title":84,"news_slug":85,"published_at":86},"cd49f913-cde7-4cf3-8d93-24508653180e","腾讯混元开源AuK:1.5B语音模型统一生成与编辑,4步推理快4.5倍","tencent-hunyuan-auk-speech-editing","2026-09-09T09:12:00+00:00"]