arXiv cs.AI· Rui Gong, Ruoyu Wang, Yalun Dai, Hao Li, Sidi Yang, Qihao Zhao, Hongyu Li, Zeyang Guan, Fangzhou Hong, Zhaoxi Chen, Kim-Hui Yap, Ziwei Liu·· 12 小时前AI 评分40
S-Agent 2.0：用通用编程智能体实现空间推理
S-Agent 2.0: Spatial Reasoning with a General Coding Agent
AI 导读
S-Agent 2.0 将现成通用编程智能体与命令行感知工具包结合，通过持久场景记忆和空间语言解读闭合证据回路。在 MMSI-Bench 和 ReVSI 上，该工具包使 Qwen3.8-27B 分别提升 13.9 和 20.5 分，Gemini-3.8-Flash 提升 11.2 和 9.6 分。
正文 · 原文
Authors:Rui Gong, Ruoyu Wang, Yalun Dai, Hao Li, Sidi Yang, Qihao Zhao, Hongyu Li, Zeyang Guan, Fangzhou Hong, Zhaoxi Chen, Kim-Hui Yap, Ziwei Liu
Abstract:Spatial reasoning asks a model to work out where objects are in 3D space, how they relate to one another, and how they move. Agentic VLMs have recently brought large improvements on such tasks by letting the model actively gather evidence with 3D perception tools instead of answering in a single pass. Even so, on many questions the tools return the right evidence and the model still does not arrive at the correct answer: the gap between model and tools, where raw readings become usable conclusions, is left unfilled. We present S-Agent 2.0, a spatial reasoning agent that equips an off-the-shelf general coding agent with a command-line perception toolkit, and we design the interface around one principle: help the VLM close the evidence loop: acquire scene evidence, remember it in a persistent scene memory, and interpret each reading in plain spatial language. This design keeps the full flexibility of a coding agent, and every reading it acts on stays grounded in the tools' geometry. On MMSI-Bench and ReVSI, the toolkit improves a strong open-source model, Qwen3.8-27B, by +13.9 and +20.5 points over the same agent without tools, and a strong closed-source model, Gemini-3.8-Flash, by +11.2 and +9.6 points. With the open backbone S-Agent 2.0 reaches 66.5 on MMSI-Bench and 82.3 on ReVSI, ahead of the specialized system SpatialClaw on the same backbone (62.6 and 74.5), and with Gemini-3.8-Flash it reaches 74.1 and 90.5 against its 67.0 and 79.7. Both tool gains shrink as the backbone grows more capable, but SpatialClaw's shrinks far more on ReVSI, from +13.2 to +2.7, while ours stays at +9.6. S-Agent 2.0 also works across different general coding agent harnesses, such as Codex and Claude Code.
| Subjects: | Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2610.28354 [cs.AI] |
| (or arXiv:2610.28354v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.28354 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Rui Gong [view email]
[v1]
Tue, 20 Oct 2026 15:20:00 UTC (8,526 KB)
来源：arXiv cs.AI · arxiv.org