killthefullmoon/PhyX
PhyX: Does Your Model Have the "Wits" for Physical Reasoning?
PhyX is a resource for evaluating how well AI models can solve complex physics problems using both text and images. It provides 3,000 university-level physics questions, complete with realistic visual scenarios. Researchers and AI developers can use this to benchmark and improve their multimodal AI models' ability to understand and reason about physical phenomena.
Use this if you are developing or evaluating AI models that need to interpret visual information alongside textual descriptions to solve challenging, real-world physics problems.
Not ideal if you are looking for a dataset focused on simple recall of physics facts or if your AI model does not process visual inputs.
Stars
52
Forks
1
Language
Python
License
MIT
Category
Last pushed
Feb 15, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/killthefullmoon/PhyX"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
MMMU-Benchmark/MMMU
This repo contains evaluation code for the paper "MMMU: A Massive Multi-discipline Multimodal...
pat-jj/DeepRetrieval
[COLM’25] DeepRetrieval — 🔥 Training Search Agent by RLVR with Retrieval Outcome
lupantech/MathVista
MathVista: data, code, and evaluation for Mathematical Reasoning in Visual Contexts
x66ccff/liveideabench
[𝐍𝐚𝐭𝐮𝐫𝐞 𝐂𝐨𝐦𝐦𝐮𝐧𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬] 🤖💡 LiveIdeaBench: Evaluating LLMs' Scientific Creativity and Idea...
ise-uiuc/magicoder
[ICML'24] Magicoder: Empowering Code Generation with OSS-Instruct