LiveBench Math Comp — leaderboard

Metric: Score. Source: livebench.ai. 73 models tracked.

Top models

#ModelScore
1Gemini 1.5 Pro (Preview 0827)54.17
2DeepSeek V2.553.12
3DeepSeek Coder V250
4GPT-4o (2024-08-06)47.92
5Claude 3.5 Sonnet (20240620)44.79
6GPT-4o ChatGPT44.79
7GPT-4 Turbo44.79
8Llama 3.1 70B Instruct41.67
9Claude 3 Opus (20240229)41.67
10GPT-4o (2024-05-13)39.58
11Qwen 2 72B Instruct37.5
12Llama 3.1 405B Instruct37.5
13Gemini 1.5 Pro (Preview 0801)37.5
14GPT-4 Preview (0125)36.46
15Gemini 1.5 Pro (0514)36.46

Interactive version: theaggregate.ai/benchmark?slug=livebench-math-comp · How the rankings work · Data refreshed daily, snapshot 2026-07-22.