What it measures

Multimodal physical world understanding benchmark across mechanics, optics, thermodynamics, and material properties.

Reading the measurement

Ask whether the evaluation rewards recalling a fact, constructing a model, running an experiment or supporting a discovery. Those tasks require different evidence. For research claims, inspect reproducibility, uncertainty and what was actually checked.

Published evaluation

Overall accuracy; PhysBench · physical-world question answering. Results retain their published native scale and task setting.

Where this fits in an evaluation suite

A correct multiple-choice answer and a supported scientific contribution demand different evidence. Check whether the score rewards a proposal, an executed experiment, or a completed research target.

Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.

A reproducible comparison

Overall accuracy; PhysBench · physical-world question answering. Results retain their published native scale and task setting.

Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.

PRIMARY EVIDENCE & FURTHER READING

  1. PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding — published result table ↗

    Chow, Wei, Mao, Jiageng, Li, Boyi, Seita, Daniel, Guizilini, Vitor, Wang, Yue · 2025

Reviewed 2026-10-26. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.