Skip to content
THE AI RACE
Leaderboard
Compare
Benchmarks
Methodology
Changelog
Movers
Time Machine
⌘K
☰
THE AI RACE
Leaderboard
Compare
Benchmarks
Methodology
Changelog
Movers
Time Machine
⌘K
☰
THE AI RACE
Leaderboard
Compare
Benchmarks
Methodology
Changelog
Movers
Time Machine
⌘K
☰
← Back to Benchmarks
FrontierMath
Frontier-level mathematics (350 problems, 4 tiers)
Reasoning & Math
Unit: %
Max: 100
Source →
Rankings (5 organizations)
1
Google DeepMind
38%
2
OpenAI
32%
3
DeepSeek
30%
4
Anthropic
22%
5
xAI
12%
Other Benchmarks in Reasoning & Math
MATH-500
AIME 2025
ARC-AGI-2
All Categories
Language & Knowledge
Coding
Reasoning & Math
Image Generation
Video Generation
Multimodal
Agents & Tools