Artificial Analysis Intelligence Index
What it measures and how
Use only the official Artificial Analysis Intelligence Index. Performance and price fields from the API do not count toward ranking.
How this evidence is used
Broad task evidence receives 30%. The 10 underlying benchmarks in v4.3 have been checked for duplicate votes with active specialized benchmarks. Keep the index intact; do not present an aggregate as an individual result.
Benchmark results
Fixed rules select one representative configuration per public model. Anonymous test models are omitted. Original source ranks are retained, with up to 30 models per benchmark.
| Source rank | Source model name | Raw result | Representative configuration |
|---|---|---|---|
| 4 | muse-wonder-fullMeta | 55.8 | Max reasoning |
| 8 | gpt-6-astraOpenAI | 52.7 | Max reasoning |
| 9 | gemini-4-argonGoogle | 52.6 | High reasoning |
| 10 | muse-wonderMeta | 52.4 | Max reasoning |
| 13 | gpt-6-1-solOpenAI | 51.8 | Max reasoning |
| 18 | claude-opus-5Anthropic | 50.8 | Max reasoning |
| 24 | muse-spark-1-3Meta | 48.1 | Max reasoning |
| 27 | gpt-6-solOpenAI | 47.6 | Max reasoning |
| 28 | gpt-5-6-solOpenAI | 47 | Max reasoning |
| 31 | grok-4-7xAI | 46.4 | xHigh reasoning |
| 32 | mimo-v2-6-proXiaomi | 46.3 | Source default configuration |
| 35 | qwen3-8-maxAlibaba | 45.4 | Source default configuration |
| 37 | glm-5-3Z.ai | 44.8 | Max reasoning |
| 40 | grok-4-6-xhighxAI | 44.2 | xHigh reasoning |
| 43 | step-5StepFun | 43.7 | Source default configuration |
| 44 | kimi-k3Moonshot AI | 43.6 | Max reasoning |
| 45 | claude-haiku-5-5Anthropic | 43.4 | Max reasoning |
| 51 | gpt-5-6-terraOpenAI | 42.1 | Max reasoning |
| 53 | claude-opus-4-8Anthropic | 41.8 | Max reasoning |
| 53 | glm-5-3-flashZ.ai | 41.8 | Source default configuration |
| 56 | ling-3-1-flashInclusionAI | 41.1 | Source default configuration |
| 57 | gemini-3-8-flashGoogle | 40.9 | High reasoning |
| 59 | claude-opus-4-7Anthropic | 40.7 | Max reasoning |
| 61 | qwen3-8-2-4t-a95bAlibaba | 39.9 | Source default configuration |
| 62 | qwen3-8-flash-nextAlibaba | 39.8 | Source default configuration |
| 65 | muse-spark-1-2Meta | 39.6 | xHigh reasoning |
| 67 | deepseek-v4-1-flashDeepSeek | 39.5 | Max reasoning |
| 70 | gemini-3-7-flashGoogle | 39.1 | High reasoning |
| 71 | gpt-5-4OpenAI | 39 | xHigh reasoning |
| 72 | muse-spark-openMeta | 38.9 | xHigh reasoning |
Limitations and attribution
Without an API key, this entire source is excluded for the round; no 50-point substitute is assigned. Web scraping must not replace the API. Display requires attribution and the index version.
Data license: API data displayed or shared must carry prominent attribution. Redistribution rights and custom terms require separate agreement and compliance with the Terms of Use.
Results are published by Artificial Analysis. Raw scores and the e/acc consensus index use different scales and cannot be added directly.