Web-Bench: A LLM Code Benchmark Based on Web Standards and Frameworks

Benchmark Model Rank Results
on-web-benchclaude-3-7-sonnet-20250219-thinking#1Pass@1: 25.11%Pass@2: 35.33%
on-web-benchclaude-3-5-sonnet-20241022#2Pass@1: 23.04%Pass@2: 32.39%
on-web-benchclaude-3-7-sonnet-20250219#3Pass@1: 22.50%Pass@2: 30.98%
on-web-benchclaude-3-5-sonnet-20240620#4Pass@1: 21.96%Pass@2: 30.33%
on-web-benchgpt-4.1#5Pass@1: 21.09%Pass@2: 25.11%
on-web-benchgpt-4.1-mini#6Pass@1: 20.76%Pass@2: 23.70%
on-web-benchdoubao-pro-1.5-thinking#7Pass@1: 20.11%Pass@2: 30.22%
on-web-benchgpt-4o#8Pass@1: 17.17%Pass@2: 23.80%
on-web-benchdeepseek-v3-0324#9Pass@1: 17.07%Pass@2: 23.59%
on-web-benchdeepseek-coder-v2#10Pass@1: 16.74%Pass@2: 23.15%
on-web-benchdoubao-pro-1.5-32k#11Pass@1: 16.63%Pass@2: 22.93%
on-web-benchllama-4 Maverick#12Pass@1: 15.98%Pass@2: 20.87%
on-web-benchqwen-max-2025-01-25#13Pass@1: 15.87%Pass@2: 19.02%
on-web-benchgemini-2.5-pro#14Pass@1: 15.67%Pass@2: 24.02%
on-web-benchclaude-3-5-haiku-20241022#15Pass@1: 15.43%Pass@2: 21.74%
on-web-benchgemini-2.0-flash#16Pass@1: 15.33%Pass@2: 20.87%
on-web-benchgemini-2.0-flash-thinking#17Pass@1: 14.89%Pass@2: 19.24%
on-web-benchgemini-pro-1.5#18Pass@1: 14.78%Pass@2: 20.87%
on-web-benchdeepseek-r1#19Pass@1: 14.46%Pass@2: 26.20%
on-web-benchstep-fun-2-16k#20Pass@1: 13.70%Pass@2: 15.87%
on-web-bencho4-mini#21Pass@1: 13.26%Pass@2: 22.93%
on-web-benchmistral-large-2411#22Pass@1: 13.04%Pass@2: 18.70%
on-web-benchgemini-flash-1.5#23Pass@1: 12.83%Pass@2: 17.07%
on-web-benchqwen-plus-2025-01-25#24Pass@1: 11.85%Pass@2: 15.11%
on-web-benchgrok-2-1212#25Pass@1: 11.30%Pass@2: 17.17%
on-web-benchqwen-2.5-72b-instruct#26Pass@1: 10.54%Pass@2: 13.70%
on-web-bencho1#27Pass@1: 10.43%Pass@2: 12.39%
on-web-benchgemma-3-27b#28Pass@1: 9.89%Pass@2: 11.85%
on-web-bencho3-mini#29Pass@1: 9.13%Pass@2: 14.24%
on-web-benchgpt-4o-mini#30Pass@1: 8.48%Pass@2: 13.04%
on-web-benchsense-chat-5#31Pass@1: 8.48%Pass@2: 12.72%
on-web-benchminimax-text#32Pass@1: 8.48%Pass@2: 10.76%
on-web-bench360-gpt2-o1#33Pass@1: 8.26%Pass@2: 14.46%
on-web-benchGLM-4-0414#34Pass@1: 7.50%Pass@2: 9.02%
on-web-benchgpt-4.1-nano#35Pass@1: 7.07%Pass@2: 12.28%
on-web-benchllama-3.3#36Pass@1: 6.63%Pass@2: 9.57%
on-web-benchmoonshot-kimi-latest#37Pass@1: 5.22%Pass@2: 11.85%
on-web-benchllama-4 Scout#38Pass@1: 5.00%Pass@2: 7.72%
on-web-benchdoubao-pro-1.5-32k-lite#39Pass@1: 3.48%Pass@2: 5.98%
on-web-benchqwen-turbo-2024-11-01#40Pass@1: 2.61%Pass@2: 5.11%