Atlas

Benchmarks

← All benchmarks

Project Euler

Math · 2025-11-28

This MathArena track evaluates models on the Project Euler problem set. Scores report the percentage of problems answered correctly.

Top models (higher is better)

ModelScore
Gemini 3.1 Pro Preview89.0
GPT-5.489.0
Opus 4.687.7
Gemini 3.5 Flash82.0
GPT-5.281.6
GLM-5.167.2
Gemini 3 Pro Preview62.4
Gemini 3 Flash Preview61.8
GPT-5.161.5
Kimi K2.560.0
GPT-556.9
DeepSeek-V3.250.2
Kimi K2 Thinking50.1
O4 Mini48.3
Grok 446.4
Grok 4 Fast46.0
Grok 4.1 Fast45.3
Gemini 2.5 Pro26.7
Loading Atlas data…