# TriPhysBench: Evaluating Physical Intelligence for Embodied Agents

DevFeed: [TriPhysBench: Evaluating Physical Intelligence for Embodied Agents](<https://devfeed.tech/articles/triphysbench-evaluating-physical-intelligence-for-embodied-agents-115238.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.01934>)

Author: Zhongxiang Sun, Chenghao Yang, Haojie Ding, Boheng Zhang, Fan Yang, Qingsong Wen, Kun Wang

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Machine Perception](<https://devfeed.tech/topics/machine-perception.md>), [ai evaluation](<https://devfeed.tech/topics/ai-evaluation.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [intelligence](<https://devfeed.tech/tags/intelligence.md>), [knowledge](<https://devfeed.tech/tags/knowledge.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [vision-language-models](<https://devfeed.tech/tags/vision-language-models.md>)

## AI overview

TriPhysBench evaluates whether embodied agents can combine physical perception, external knowledge retrieval, and quantitative reasoning. It contains 410 questions drawn from 200 videos, split into measurement-grounded and knowledge-grounded subtasks that require gathering observations or external specifications and applying them to quantitative problems. Across six vision-language models, the best answer accuracy is 50.38%. The authors also present TriPhys, which trains a unified policy on tool-integrated trajectories, using intermediate reasoning to guide retrieval and computation and tool feedback to inform later inference. On 110 questions from videos excluded from training, it improves answer and solution scores across three Qwen configurations; TriPhys-27B raises answer accuracy by 17.7% relative to its Qwen3.6-27B backbone.

## Source excerpt

arXiv:2611.01934v1 Announce Type: new Abstract: Embodied agents need physical intelligence to understand their surroundings through physical perception, physical knowledge retrieval, and physical reasoning. These capabilities must work together to interpret events and make quantitative judgments. Existing benchmarks advance physical understanding and general-purpose tool use, yet the coordination of these capabilities in physical scenarios remains insufficiently characterized. Inspired by Popper's Three Worlds, we introduce TriPhysBench, a benchmark of 410 questions from 200 videos that brings physical perception, physical knowledge retrieval, and physical reasoning into a shared evaluation setting. Its measurement-grounded and knowledge-grounded subtasks impose complementary demands on obtaining observations and acquiring external specifications, then applying them to quantitative problems. Among six representative vision-language models, the best achieves an answer accuracy of only 50.38%, revealing limited performance on tasks that require these capabilities to work together. We further introduce TriPhys, a framework that learns a unified policy for physical perception, knowledge retrieval, and quantitative reasoning from complete tool-integrated trajectories. The policy uses intermediate reasoning to guide retrieval and computation, then incorporates tool feedback into subsequent physical inference. On 110 questions from videos excluded from training, TriPhys improves answer and solution scores across three Qwen configurations, with a 17.7% relative gain in answer accuracy for TriPhys-27B over its Qwen3.6-27B backbone. This work provides a benchmark and a learning pathway toward embodied agents that move beyond recognizing events to understanding and reasoning about the physical processes behind them.