Benchmark Run Explorer

Explore individual benchmark runs, view score distributions, and analyze performance across models and providers.

Last updated: 4/30/2026, 5:53:02 PM

1000 runs•284 models

Model Rankings

#1 Claude Opus 4.1
2 providers
Score: 61.5%
Range: 52-77%
4 runs

#2 GPT-5.4 Low OpenAI
Score: 76.5%
1 run

#3 Kimi 2.6
2 providers
Score: 60.6%
Range: 46-76%
3 runs

#4 Claude Opus 4.5
3 providers
Score: 72.1%
1 run

#5 Z.AI GLM-5 Z.AI
Score: 70.0%
Range: 21-80%
14 runs

#6 Claude Sonnet 4.5
4 providers
Score: 70.0%
Range: 17-77%
7 runs

#7 Kimi For-Coding Moonshot
Score: 72.9%
1 run

#8 GPT-5.1 Medium
2 providers
Score: 63.5%
Range: 55-80%
7 runs

#9 Claude Opus 4.5 Thinking
Unknown
Score: 72.1%
1 run

#10 Claude 3.5 Sonnet Custom
Score: 71.5%
1 run

#11 GPT-5.3 Codex Medium Codex CLI
Score: 71.5%
Range: 53-71%
6 runs

#12 GPT-5.5 Medium Codex CLI
Score: 71.2%
Range: 70-71%
2 runs

#13 pony-alpha OpenRouter
Score: 70.9%
Range: 64-75%
3 runs

#14 GPT-5.5 High Codex CLI
Score: 70.9%
Range: 70-71%
2 runs

#15 GPT-5.5 Xhigh Codex CLI
Score: 70.9%
Range: 70-71%
2 runs

#16 doubao-seed-2.0-code Custom
Score: 70.6%
1 run

#17 GPT-5.4 OpenAI
Score: 70.0%
Range: 50-70%
3 runs

#18 GPT-5.3 Codex Xhigh Codex CLI
Score: 70.0%
Range: 70-70%
2 runs

#19 GPT-5.3 Codex High Codex CLI
Score: 70.0%
Range: 66-70%
3 runs

#20 Gemini 3.1 Pro Preview
2 providers
Score: 70.0%
1 run