Skip to content
Agents tracked: 502 Releases (7d): 613 Agent status: 4 with issuesSince last limit reset: Claude 7d · Codex 5d For AI agents: Markdown Updated Nov 18, 2026

Claude Opus 5.5 is best at coding with 71.1, 4.2 ahead of Claude Sonnet 5.5

Coding: average of 3 tests (MirrorCode, FrontierSWE v2, WeirdML v3), 0 to 100. Live chart from AgentGid, updated daily.

Claude Opus 5.5 is best at coding with 71.1, 4.2 ahead of Claude Sonnet 5.5

Coding: average of 3 tests (MirrorCode, FrontierSWE v2, WeirdML v3), 0 to 100

0255075100Claude Opus 5.5 (Anthropic): 71.1Claude Opus 5.571.1Claude Sonnet 5.5 (Anthropic): 66.9Claude Sonnet 5.566.9Claude Fable 5.1 (Anthropic): 66.0Claude Fable 5.166.0GPT-6 Astra (OpenAI): 61.9GPT-6 Astra61.9GPT-6.1 Terra (OpenAI): 60.8GPT-6.1 Terra60.8 · 1 test fewerGPT-6.1 Sol (OpenAI): 58.7GPT-6.1 Sol58.7 · 1 test fewerClaude Opus 5 (Anthropic): 58.6Claude Opus 558.6Gemini 4 Argon (Google): 53.3Gemini 4 Argon53.3Claude Haiku 5.5 (Anthropic): 49.8Claude Haiku 5.549.8Claude Fable 5 (Anthropic): 49.3Claude Fable 549.3GPT-6 Sol (OpenAI): 39.8GPT-6 Sol39.8 · 1 test fewerDeepSeek V4.1 Flash (DeepSeek): 38.7DeepSeek V4.1 Flash38.7 · 1 test fewer
Each model's best effort level. Models need 2 tests in this group to be listed; less than 1 point apart is a tie (=). Source: Gid Bench by AgentGid from Epoch AI (CC BY 4.0) and makers' launch pages · Nov 17, 2026.

Put this chart on your site

The chart updates by itself. Please keep the link to the source.

More charts