PHYRD benchmark tests LLMs on reproducing and extending physics
Across five models, Reproduce accuracy was 49.21%–68.25% and closed-book Discover was 18.60%–74.42%.
48
AI model
Qwen-35B-A3B
- Mentions · 7d0
- Mentions · 30d1
- All mentions2
- First seenOct 12, 2026
- Last mentioned19d ago
Timeline
2 articles · newest firstCowork-Gym: Synthesizing Interactive Work Environments at Scale
Cowork-Gym post-trains Qwen agents on 64.9K synthetic workplaces, lifting six-benchmark scores.
58
Appears with
- OpenAIorganization · 1
- FormulaForestproduct · 1
- Qwen-9Bai model · 1
- PHYRDproduct · 1
- FrontierScienceproduct · 1
- Qwen-27Bai model · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.