Claude Opus 5 Tops MathArena Evaluation with Over 90% Accuracy
scaling01 · x · 2026-07-31
MathArena evaluated Claude Opus 5. The model secured 4th and 2nd place on ArXivMath and ArXivLean respectively. Notably, it crushed the BrokenArXiv benchmark, reaching 1st place with over 90% accuracy.
More from Models
- Claude Keeps Hallucinating Itself as 'A Helpful Assistant Hovering Above' — Sauers_ · 2026-07-31
- Does Claude's Mood Affect Reward Hacking? Community Calls for Specific Evals — 1a3orn · 2026-07-31
- Kimi K3 Beats Opus 4.8 in 34-Prompt Oneshot Eval at 1/16 the Cost — kms_dev · 2026-07-31
- Llama 3.1 405B Cached Input Price Drops to $0.02/M Tokens — gabrielchua · 2026-07-31
- Thinking Machines Launches Inkling-Small: 12B Active Params Beats Trillion-Param Model — 新智元 · 2026-07-31
- xAI Launches SuperGrok Plus at $100/mo, Mid-Tier Between $30 and $300 Plans — XFreeze · 2026-07-31