Benchmarks / External Composite Indices
AA Intelligence Index
Artificial Analysis's composite intelligence index. Displayed for reference; excluded from the BenchAtlas Index (its inputs already feed it). Comparable only within a major index version.
maintained by Artificial Analysis · 243 observations on record
Versions
1 version
| Version | Tasks |
|---|---|
| Intelligence Index v4.1 current aa-intelligence-index-v4-1 | — |
Current leaderboard
Intelligence Index v4.1 · Index points · top 25 of 243 systems
| # | System | Index points |
|---|---|---|
| 1 | GPT 5.6 Solmax effort OpenAI | 58.9 points |
| 2 | GPT 5.6 Solxhigh effort OpenAI | 57.7 points |
| 3 | GPT 5.6 Solhigh effort OpenAI | 55.9 points |
| 4 | Claude Opus 4.8max effort Anthropic | 55.7 points |
| 5 | GPT 5.6 Terramax effort OpenAI | 55.0 points |
| 6 | GPT 5.5 (2026-04-22)xhigh effort OpenAI | 54.8 points |
| 7 | Grok 4.5high effort xAI | 53.8 points |
| 8 | GPT 5.6 Solmedium effort OpenAI | 53.6 points |
| 9 | Claude Opus 4.7max effort Anthropic | 53.5 points |
| 10 | Claude Sonnet 5max effort Anthropic | 53.4 points |
| 11 | GPT 5.5 (2026-04-22)high effort OpenAI | 53.1 points |
| 12 | GPT 5.6 Terraxhigh effort OpenAI | 51.6 points |
| 13 | GPT 5.4 (2026-03-05)xhigh effort OpenAI | 51.4 points |
| 14 | GPT 5.6 Lunamax effort OpenAI | 51.2 points |
| 15 | GLM 5.2max effort Z.ai | 51.1 points |
| 16 | Muse Spark 1.1xhigh effort Meta AI | 50.6 points |
| 17 | GPT 5.5 (2026-04-22)medium effort OpenAI | 50.4 points |
| 18 | Gemini 3.5 Flashhigh effort Google DeepMind | 50.2 points |
| 19 | GPT 5.6 Sollow effort OpenAI | 49.4 points |
| 20 | GPT 5.6 Lunaxhigh effort OpenAI | 49.1 points |
| 21 | GPT 5.6 Terrahigh effort OpenAI | 49.0 points |
| 22 | Claude Sonnet 4.6max effort Anthropic | 47.2 points |
| 23 | Gemini 3.1 Pro Preview Google DeepMind | 46.5 points |
| 24 | GPT 5.6 Lunahigh effort OpenAI | 46.1 points |
| 25 | Qwen3 7 Max Alibaba (Qwen) | 46.0 points |
One row per evaluated system — reasoning-effort variants rank separately.