• New Chat
  • Leaderboard
  • Take your chats anywhere

    Create an account to save your chat history across your devices.

    • Terms of UsePrivacy Policy
      OverviewTextWebDevVisionText-to-ImageImage EditSearchText-to-VideoImage-to-VideoCopilot
      OverviewTextWebDevVisionText-to-ImageImage EditSearchText-to-VideoImage-to-VideoCopilot

      Copilot Arena

      Compare how well AI coding assistants understand and generate code across various programming languages and tasks.
      Learn more about Copilot Arena at our blog.

      Last Updated

      May 29, 2025

      Total Votes

      32,101

      Total Models

      14

      /
      /
      1
      Anthropic
      Claude 3.5 Sonnet (06/20)
      1012+11/-123,544
      Anthropic
      Proprietary
      1
      Deepseek V2.5 (FIM)
      1028+15/-162,292
      Deepseek AI
      Deepseek
      1
      Anthropic
      Claude 3.5 Sonnet (10/22)
      1004+12/-103,596
      Anthropic
      Proprietary
      1
      Codestral (25.01)
      1001+13/-122,180
      Mistral
      Mistral AI Non-Production
      1
      Mercury Coder Mini
      994+19/-181,430
      Inception AI
      Unreleased
      1
      Qwen Icon
      Qwen-2.5-Coder (FiM)
      998+16/-133,401
      Alibaba
      Alibaba
      2
      Codestral (05/24)
      1001+6/-95,744
      Mistral
      Mistral AI Non-Production
      3
      GPT-4o (08/06)
      986+11/-104,464
      OpenAI
      Proprietary
      3
      Gemini-1.5-Pro-002
      986+10/-113,441
      Google
      Proprietary
      4
      Meta
      Meta-Llama-3.1-405B-Instruct
      984+10/-103,432
      Meta
      Llama 3.1 Community
      5
      Gemini-1.5-Flash-002
      977+11/-115,191
      Google
      Proprietary
      7
      Meta
      Meta-Llama-3.1-70B-Instruct
      969+11/-74,085
      Meta
      Llama 3.1 Community
      13
      GPT-4o-mini (07/18)
      943+10/-124,776
      OpenAI
      Proprietary
      13
      Qwen Icon
      Qwen2.5-Coder-32B-Instruct
      950+8/-124,400
      Alibaba
      Apache 2.0