# OVO-S-Bench：面向多模态大语言模型流式空间智能的分层基准

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-06-02 08:00
- AIHOT 分数：64
- AIHOT 链接：https://aihot.virxact.com/items/cmpz4oprn008tslkpef4upy5a
- 原文链接：https://arxiv.org/abs/2606.03890

## AI 摘要

OVO-S-Bench是一个完全人工标注的流式空间智能基准，包含1,680个问题，覆盖348个源视频。标注由12名标注员经过约804人小时的多轮质量审核完成。每个问题带有查询时间戳与证据区间，模型仅能看到查询前的视频前缀。问题分为四个抽象层级：瞬时自我中心感知、时空上下文追踪、空间模拟与推理、以及全中心映射。在38个开源与闭源MLLM上，Gemini-3.1-Pro得分59.2，落后人类专家（86.6）27个百分点，全中心映射是主要瓶颈。流式与空间微调MLLM的表现不及它们的基础模型，且链式推理会在缺乏流式依据时放大空间错误。

## 正文

Multimodal agents in robotics, AR, and autonomous driving must reason about places and layouts from continuous egocentric streams, often using evidence outside the current view. Existing benchmarks either evaluate offline over full videos or target events rather than spatial structure. We introduce OVO-S-Bench, a fully human-annotated benchmark for streaming spatial intelligence, comprising 1,680 questions over 348 source videos. Annotation involves 12 trained annotators, each also serving as a blind cross-reviewer, across roughly 804 person-hours of multi-round quality assurance. Each question carries a query timestamp and an evidence interval, and at evaluation, the model sees only the prefix preceding the query. Questions span four levels of increasing abstraction: instantaneous egocentric perception, spatiotemporal context tracking, spatial simulation and reasoning, and allocentric mapping. Across 38 proprietary and open-source MLLMs, Gemini-3.1-Pro trails human experts by 27 points, 59.2 vs. 86.6, with allocentric mapping as the dominant bottleneck. Notably, streaming and spatially fine-tuned MLLMs underperform their own backbones. We further find that chain-of-thought reasoning amplifies spatial errors when ungrounded in the stream. By exposing these limitations, OVO-S-Bench establishes a demanding testbed for next-generation streaming spatial MLLMs.