AI Ranking.
Intelligence, code, agents.
Find the model for your next idea.
Intelligence index
86 modelsArtificial Analysis Intelligence Index. Highest scores first; models without evaluations are excluded.
| # | Model / Creator | AA Index | Input / 1M | Output / 1M | Context |
|---|---|---|---|---|---|
| 01 | Claude Fable 5.1anthropic | 53.4 | $10.00 | $50.00 | 1M |
| 02 | GPT-6 Astraopenai | 52.7 | $10.00 | $50.00 | 1.1M |
| 03 | Claude Opus 5anthropic | 50.8 | $5.00 | $25.00 | 1M |
| 04 | Claude Fable 5anthropic | 49.6 | $10.00 | $50.00 | 1M |
| 05 | GPT-5.6 Solopenai | 47 | $2.00 | $10.00 | 1.1M |
| 06 | Qwen3.8 Max (0902)qwen | 45.4 | $2.00 | $6.00 | 1M |
| 07 | GLM 5.3z-ai | 44.8 | $0.91 | $2.86 | 1.3M |
| 08 | Grok 4.6x-ai | 44.3 | $2.00 | $6.00 | 500K |
| 09 | Kimi K3moonshotai | 43.6 | $1.70 | $8.50 | 1M |
| 10 | GPT-5.6 Terraopenai | 42.1 | $2.00 | $12.00 | 1.1M |
| 11 | Claude Opus 4.8anthropic | 41.8 | $5.00 | $25.00 | 1M |
| 11 | GLM 5.3 Flashz-ai | 41.8 | $0.09 | $0.30 | 1.3M |
| 13 | Gemini 3.8 Flashgoogle | 40.9 | $0.75 | $3.75 | 1M |
| 14 | Qwen3.8 2.4T A95Bqwen | 39.9 | $2.00 | $6.00 | 1M |
| 15 | Muse Spark 1.2meta | 39.6 | $1.25 | $4.25 | 1M |
| 16 | DeepSeek V4.1 Flashdeepseek | 39.5 | $0.15 | $0.60 | 1M |
| 17 | Gemini 3.7 Flashgoogle | 39.1 | $0.75 | $3.75 | 1M |
| 18 | Grok 4.5x-ai | 38.8 | $2.00 | $6.00 | 500K |
| 19 | GPT-5.5openai | 38.4 | $5.00 | $30.00 | 1.1M |
| 20 | Claude Sonnet 5anthropic | 38.2 | $2.00 | $10.00 | 1M |
| 21 | GPT-5.6 Lunaopenai | 37.3 | $0.20 | $1.20 | 1.1M |
| 22 | DeepSeek V4 Pro 0813deepseek | 36 | $0.66 | $1.98 | 1M |
| 23 | DeepSeek V4 Flash 0731deepseek | 34.3 | $0.04 | $0.16 | 1.3M |
| 24 | Gemini 3.6 Flashgoogle | 34 | $0.75 | $3.75 | 1M |
| 25 | Muse Spark 1.1meta | 33.7 | $1.25 | $4.25 | 1M |
Behind the ranking Sources, criteria and comparison limits
Indices are from Artificial Analysis, published in the catalog by OpenRouter. Each tab uses its own metric. Scores are indices, not accuracy percentages or a guarantee of results.
The catalog includes models with text output. This list does not measure tool popularity or image and video generation quality. Models missing the selected index are omitted; missing values never become zero. Access variants of the same model are grouped, using the base route when available. Ties share a position.
Prices are in US dollars per million tokens, excluding taxes, caching, tools and additional services. They may vary by provider. The displayed date is when the catalog was retrieved; the API does not report each benchmark's measurement date or version.
Cost sorting includes only models with text output and no fixed per-request fee, keeping media pricing separate from token pricing.
Public data is available without an API key. It refreshes when you open this page and every five minutes while visible. If the source is unavailable, the last copy remains with its date.