Rabdos Launches Math Reasoning Index, Claude Opus 5 Takes the Lead

prof_g · x · 2026-08-06

Rabdos AI has introduced the Rabdos Math Index, a holistic measurement designed to evaluate the mathematical reasoning capabilities of LLMs across proof, solution, and perception.

Anthropic's Claude Opus 5 currently leads the index with a score of 47, followed by OpenAI's GPT-5.6 Sol at 41. No other evaluated model managed to score above 25.

Original post →

More from Models

Models channel →