Benchmark Explorer

Explore how models perform on various benchmarks

Benchmarks

🔍

LiveBench (Coding)

Assesses performance on LiveBench coding tasks from Leetcode and AtCoder, including both code generation and code completion challenges.
Source:

Model Performance

#1
OpenAI-O4-mini-high
80.0%
#2
Claude-4.0-Sonnet
78.2%
#3
OpenAI-O3-medium
77.9%
#4
GPT-4o-0513
77.5%
#5
OpenAI-O3-high
76.7%
#6
GPT-4.5
76.1%
#7
GPT-5
75.3%
#8
Claude-3.7-Sonnet
74.3%
#9
OpenAI-O4-mini-medium
74.2%
#10
Claude-4.1-Opus-Thinking
74.0%
#11
Claude-3.5-Sonnet-1022
73.9%
#12
Claude-4.0-Sonnet-Thinking
73.6%
#13
Claude-4.0-Opus
73.6%
#14
Claude-4.0-Opus-Thinking
73.2%
#15
GPT-4.1
73.2%
#16
Claude-3.7-Sonnet-Thinking
73.2%
#17
GPT-5-mini
72.9%
#18
Kimi-K2-Instruct
71.8%
#19
DeepSeek-R1
71.4%
#20
GPT-5-nano
70.8%
#21
Gemini-2.5-Pro-0325
70.7%
#22
DeepSeek-V3
68.5%
#23
Qwen-3
64.2%
#24
Gemini-2.5-Flash
63.5%
#25
Mistral-Large
62.9%
#26
Mistral-Medium-3.1
61.5%
#27
Phi-4
60.6%
#28
Gemini-2.5-Flash-Thinking
59.2%
#29
GPT-OSS-120B
58.8%
#30
Grok-3-Mini-Beta
54.5%
#31
Cohere-Command-A
54.3%
#32
Llama-4-Maverick-17B
54.2%
#33
Claude-3.5-Haiku
53.2%
#34
Cohere-Command-R-Plus
27.1%