Los mejores modelos de IA en la actualidad
Clasificados por pruebas independientes, con lo que cuesta cada uno. Compara dos de ellos lado a lado.
| # | Modelo | Fabricante | Índice de inteligencia | Velocidad (tokens/s) | Entrada (1M tokens) | Salida (1M tokens) |
|---|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 (Max, Fallback por defecto) | Anthropic | 57.6 | 97 | $4 | $20 |
| 2 | Claude Sonnet 5.5 (Max, Fallback por defecto) | Anthropic | 56 | 141 | $2 | $10 |
| 3 | Claude Opus 5.5 (Xhigh, Fallback por defecto) | Anthropic | 56 | 77 | $4 | $20 |
| 4 | Muse Wonder Full (Max) | Meta | 55.8 | – | n/a | n/a |
| 5 | Kimi K3.1 Pro | Kimi | 55.1 | 38 | $2 | $10 |
| 6 | Claude Opus 5.5 (High, Fallback por defecto) | Anthropic | 53.6 | 74 | $4 | $20 |
| 7 | Claude Fable 5.1 (Max, Fallback por defecto) | Anthropic | 53.4 | 66 | $10 | $50 |
| 8 | Claude Fable 5.1 (Xhigh, Fallback por defecto) | Anthropic | 53.2 | 56 | $10 | $50 |
| 9 | GPT-6 Astra (Max) | OpenAI | 52.7 | 64 | $10 | $50 |
| 10 | Gemini 4 Argon (High) | 52.6 | 80 | $2 | $10 | |
| 11 | Muse Wonder (Max) | Meta | 52.4 | 158 | $4 | $20 |
| 12 | GPT-6 Astra (Xhigh) | OpenAI | 52.4 | 59 | $10 | $50 |
| 13 | Claude Sonnet 5.5 (Xhigh, Fallback por defecto) | Anthropic | 51.9 | 98 | $2 | $10 |
| 14 | GPT-6.1 Sol (Max) | OpenAI | 51.8 | 66 | $2 | $10 |
| 15 | Claude Opus 5.5 (Medium, Fallback por defecto) | Anthropic | 51.2 | 79 | $4 | $20 |
| 16 | Claude Fable 5.1 (High, Fallback por defecto) | Anthropic | 51.2 | 52 | $10 | $50 |
| 17 | GPT-6.1 Sol (Xhigh) | OpenAI | 51 | 65 | $2 | $10 |
| 18 | GPT-6 Astra (High) | OpenAI | 50.9 | 53 | $10 | $50 |
| 19 | Claude Opus 5 (Max) | Anthropic | 50.8 | – | $5 | $25 |
| 20 | GPT-6.1 Sol (High) | OpenAI | 50.2 | 60 | $2 | $10 |
| 21 | Claude Opus 5 (Xhigh) | Anthropic | 49.7 | – | $5 | $25 |
| 22 | Claude Fable 5 (Max, Fallback de Opus 4.8) | Anthropic | 49.6 | – | $10 | $50 |
| 23 | GPT-6 Astra (Medium) | OpenAI | 49.6 | 49 | $10 | $50 |
| 24 | Claude Fable 5.1 (Medium, Fallback por defecto) | Anthropic | 48.9 | 50 | $10 | $50 |
| 25 | Claude Opus 5 (High) | Anthropic | 48.1 | – | $5 | $25 |
| 26 | GPT-6.1 Terra (Max) | OpenAI | 48.1 | 91 | $1 | $5 |
| 27 | Muse Spark 1.3 (Max) | Meta | 48.1 | – | $1.25 | $4.25 |
| 28 | GPT-6.1 Sol (Medium) | OpenAI | 47.8 | 56 | $2 | $10 |
| 29 | GPT-6 Sol (Max) | OpenAI | 47.6 | – | $2 | $10 |
| 30 | GPT-5.6 Sol (Max) | OpenAI | 47 | – | $4 | $20 |
| 31 | Claude Fable 5.1 (Low, Fallback por defecto) | Anthropic | 46.8 | 52 | $10 | $50 |
| 32 | Claude Sonnet 5.5 (High, Fallback por defecto) | Anthropic | 46.8 | 101 | $2 | $10 |
| 33 | Grok 4.7 (Xhigh) | SpaceXAI | 46.4 | 77 | $2 | $6 |
| 34 | Grok 4.7 (High) | SpaceXAI | 46.3 | 84 | $2 | $6 |
| 35 | MiMo-V2.6-Pro | Xiaomi | 46.3 | 47 | $0.43 | $0.87 |
| 36 | GPT-6 Astra (Low) | OpenAI | 45.8 | 57 | $10 | $50 |
| 37 | Qwen3.8 Max (0902) | Alibaba | 45.4 | 40 | $2 | $6 |
| 38 | GPT-6.1 Terra (Xhigh) | OpenAI | 45.4 | 91 | $1 | $5 |
| 39 | DeepSeek V4.1 Flash 1017 (Max) | DeepSeek | 45.1 | 159 | $0.30 | $1.20 |
| 40 | Muse Spark 1.3 (Xhigh) | Meta | 45.1 | – | $1.25 | $4.25 |
Datos procedentes de Artificial Analysis, actualizados por última vez el 27 de octubre de 2026 a las 15:17 UTC. El valor n/a significa que el modelo carece de un precio público (suele ocurrir cuando no dispone de API pública): nunca se considera gratuito. El índice de inteligencia combina varios tests independientes. Los precios se expresan en dólares estadounidenses por cada millón de tokens: input es el texto que tú envías al modelo, output es el texto que él te devuelve. Un token equivale a un fragmento pequeño de texto, de aproximadamente tres cuartas partes de una palabra.