Benchmark Explorer

Explore how models perform on various benchmarks

Benchmarks

🔍

SWE-bench

Evaluates the model's ability to resolve real-world software engineering issues from GitHub repositories. Models are tested on their capacity to generate code patches that fix actual bugs and implement requested features from open-source projects.
Source:

Model Performance

#1
Claude-4.0-Sonnet
65.0%
#2
Grok-4
58.6%
#3
Grok-4-Thinking
58.6%
#4
OpenAI-O3-high
49.8%
#5
OpenAI-O3-medium
49.8%
#6
GPT-4.1
47.4%
#7
Qwen-3-Max-Preview
44.4%
#8
Grok-3-Beta
42.0%
#9
Grok-3-Think
42.0%
#10
Kimi-K2-Instruct
34.2%
#11
OpenAI-O4-mini-medium
33.4%
#12
OpenAI-O4-mini-high
33.4%
#13
Llama-4-Maverick-17B
18.4%
#14
Cohere-Command-A
0.2%