EvalLedger
Leaderboard
Compare
Token Calculator
Model Releases
Version series
Changelog
EN
Benchmark
MathArena Apex
Hard competition-mathematics problems.
Score history →
Model scores
#
Model
Score
1
DeepSeek V4.1-Flash 0910
65.6%
2
Kimi K3
65.6%
3
DeepSeek V4-Pro 0813
65.3%
4
Gemini 3.1 Pro
60.9%
5
DeepSeek V4-Flash 0731
58.6%