Benchmark Explorer

Explore how models perform on various benchmarks

Benchmarks

🔍

Math500

Evaluates the model's accuracy on a set of 500 advanced mathematics problems, requiring a high level of mathematical reasoning and problem-solving skills.
Source:

Model Performance

#1
Grok-4
96.2%
#2
Grok-4-Thinking
96.2%
#3
Claude-4.1-Opus-Thinking
95.4%
#4
Gemini-2.5-Pro-Thinking
95.2%
#5
Gemini-2.5-Pro-0325
95.2%
#6
GPT-5-mini
94.8%
#7
OpenAI-O3-high
94.6%
#8
OpenAI-O3-medium
94.6%
#9
Qwen-3
94.6%
#10
Qwen-3-Thinking
94.6%
#11
OpenAI-O4-mini-medium
94.2%
#12
Grok-3-Mini-Beta
94.2%
#13
OpenAI-O4-mini-high
94.2%
#14
Kimi-K2-Instruct
94.2%
#15
Claude-4.0-Sonnet-Thinking
93.8%
#16
Claude-4.1-Opus
93.0%
#17
DeepSeek-R1
92.2%
#18
OpenAI-O3-mini-high
91.8%
#19
Gemini-2.5-Flash-Thinking
91.8%
#20
OpenAI-O3-mini-medium
91.8%
#21
Gemini-2.5-Flash
91.6%
#22
Claude-3.7-Sonnet-Thinking
91.6%
#23
OpenAI-O1-1217
90.4%
#24
Claude-4.0-Opus
90.4%
#25
Claude-4.0-Sonnet
90.3%
#26
Grok-3-Beta
89.8%
#27
Grok-3-Think
89.8%
#28
DeepSeek-V3-0324
88.6%
#29
Gemini-2.0-Flash
88.0%
#30
GPT-4.1
87.2%
#31
Mistral-Medium-3.1
87.0%
#32
Llama-4-Maverick-17B
85.2%
#33
Gemini-2.0-Flash-Thinking-0121
84.6%
#34
Gemini-1.5-Pro
82.8%
#35
DeepSeek-V3
80.4%
#36
Claude-3.7-Sonnet
76.8%
#37
Cohere-Command-A
76.2%
#38
Mistral-Large-2
74.4%
#39
Llama-3.3-70B
73.4%
#40
GPT-4-mini
72.6%
#41
Claude-3.5-Sonnet-1022
72.4%
#42
Llama-3.1-405B
71.4%
#43
Claude-3.5-Haiku
64.2%