Expert re-grading shows physics benchmarks are broken: most 'model errors' are benchmark or grading errors
zainhas · x · 2026-09-16
An arXiv paper with 50+ authors (including physics professors) revisits frontier LLM performance on six widely used physics benchmarks. Low reported scores — including those in the Artificial Analysis Intelligence Index (2026) — suggest models struggle with advanced physics, but expert re-grading reveals that most errors are benchmark errors or grading errors, not model errors. Several benchmarks are near saturation and can no longer reliably distinguish frontier models.
More from Models
- 'You Can Feel the Gradient Slopes': Dev Contrasts Benchmaxxed RL Models vs General Ones — theteknosaur · 2026-11-02
- LightOnOCR-3 Launches: Tops FRBench, Adds Layout Structure and Chart Extraction — IgorCarron · 2026-11-02
- NVIDIA's Kumo Is a Family of Four Tabular Foundation Models, Not Just One — pandeyparul · 2026-11-02
- SemiAnalysis: safety classifiers are underrated — Meta's Muse tops cyber benchmark at 35.2% — ThomasScialom · 2026-11-02
- DeepSeek V4.1 Flash 1017 wins over a longtime closed-source-only user: fast, cheap, reliable — _bachrc · 2026-11-02
- Local quants vs Claude on 9 real private bugs: 3-bit costs thinking time, not score — FeydRowan · 2026-11-02