| GPT-5.5-So (OpenAI) | 80 score / USD | 估读 |
| Bar Chart: Composite average pass@1 | Codex - GPT-5.5 (OpenAI) | 77 score / USD | 图上印着 |
| Bar Chart: Composite average pass@1 | Claude Code - Fable 5 (with fallback) (Anthropic) | 77 score / USD | 图上印着 |
| GPT-5.5 (OpenAI) | 76 score / USD | 估读 |
| Bar Chart: Composite average pass@1 | Grok Build - Grok 4.5 (high) (SpaceXAI) | 76 score / USD | 图上印着 |
| Bar Chart: Composite average pass@1 | Codex - GPT-5.5-Luna (OpenAI) | 75 score / USD | 图上印着 |
| Bar Chart: Composite average pass@1 | Claude Code - Opus 4.8 (Anthropic) | 73 score / USD | 图上印着 |
| GPT-5.5 (OpenAI) | 71 score / USD | 估读 |
| Scatter Plot: Coding Agent Index vs. Cost per Task | Codex - GPT-5.6 Luna (high) (OpenAI) | 69 score / USD | 估读 |
| Bar Chart: Composite average pass@1 | Claude Code - Opus 4.8 (high) (Anthropic) | 67 score / USD | 图上印着 |
| Bar Chart: Composite average pass@1 | GLM-5.2 (Novita) | 58 score / USD | 估读 |
| Scatter Plot: Coding Agent Index vs. Cost per Task | Cursor CLI - Composer 2.5 (Cursor) | 52 score / USD | 估读 |