Atlas

Benchmarks

← All benchmarks

ArXivMath 04/2026

Math · 2026-05-01

This MathArena track evaluates models on the ArXivMath 04/2026 problem set. Scores report the percentage of problems answered correctly.

Top models (higher is better)

ModelScore
Claude Fable 570.7
GPT-5.567.1
Gemini 3.1 Pro Preview62.2
Opus 4.861.0
Opus 4.758.5
Kimi K353.7
DeepSeek-V4-Pro52.4
Gemini 3.5 Flash51.2
Gemini 3.6 Flash51.2
DeepSeek-V4-Flash48.8
GLM-5.248.0
Step 3.7 Flash37.4
GLM-5.136.6
Qwen3.6 35B A3B34.1
Step 3.5 Flash34.1
Qwen3.5-2B2.4
Loading Atlas data…