Artificial Analysis Coding Index aa-coding-index Leaderboard
AA's cross-model coding aggregate (LiveCodeBench + SciCode + more) — the coding-focused sibling of the Intelligence Index. Live from the AA API. · Metric: Coding Index (higher is better) · 🟢 Updated 10h ago
| # | Model | Coding Index | Paper |
|---|---|---|---|
| 1 | GPT-5.6 Sol (xhigh) | 78.30 | — |
| 2 | GPT-5.6 Sol (max) | 77.40 | — |
| 3 | GPT-5.6 Sol (high) | 77.20 | — |
| 4 | GPT-5.6 Terra (max) | 76.70 | — |
| 5 | Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | 76.50 | — |
| 6 | GPT-5.6 Sol (medium) | 76.30 | — |
| 7 | Kimi K3 | 76.20 | — |
| 8 | GPT-5.5 (xhigh) | 74.90 | — |
| 9 | Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | 74.30 | — |
| 10 | Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | 73.60 | — |
| 11 | Grok 4.5 (high) | 72.40 | — |
| 12 | GPT-5.5 (high) | 71.60 | — |
| 13 | Claude Sonnet 5 (Adaptive Reasoning, Max Effort) | 71.50 | — |
| 14 | GPT-5.5 (medium) | 71.50 | — |
| 15 | GPT-5.6 Luna (max) | 71.40 | — |
| 16 | Muse Spark 1.1 (xhigh) | 71.30 | — |
| 17 | GPT-5.4 (xhigh) | 71.10 | — |
| 18 | GPT-5.6 Terra (xhigh) | 70.60 | — |
| 19 | Gemini 3.5 Flash (high) | 70.10 | — |
| 20 | GPT-5.6 Sol (low) | 69.70 | — |
| 21 | Gemini 3.6 Flash (high) | 69.20 | — |
| 22 | Gemini 3.1 Pro Preview | 68.80 | — |
| 23 | GLM-5.2 (max) | 68.80 | — |
| 24 | GPT-5.6 Luna (xhigh) | 68.60 | — |
| 25 | GPT-5.6 Terra (high) | 67.10 | — |
| 26 | Claude Sonnet 5 (Non-reasoning, High Effort) | 66.40 | — |
| 27 | Qwen3.7 Max | 66.00 | — |
| 28 | GPT-5.6 Sol (Non-reasoning) | 65.10 | — |
| 29 | GPT-5.6 Terra (medium) | 64.70 | — |
| 30 | GPT-5.6 Luna (high) | 63.30 | — |