Ranking de modelos de inteligencia artificial.
Índices de Artificial Analysis — inteligencia, código y agentic (0–100) — obtenidos vía la API pública de OpenRouter.
Inteligencia
Capacidad general de razonamiento, comprensión y precisión en tareas complejas. El índice principal del ranking.
Código
Habilidad para generar, depurar y optimizar código fuente en distintos lenguajes de programación.
Agente
Capacidad para ejecutar tareas autónomas: seguir instrucciones, usar herramientas y razonar en varios pasos.
Precio
Costo por millón de tokens de salida. En el radar precio vs. inteligencia, arriba a la izquierda es más inteligencia por dólar.
Inteligencia
Código
Agentic
Líderes: Inteligencia Anthropic: Claude Fable 5.1, Código Anthropic: Claude Fable 5.1, Agentic Anthropic: Claude Fable 5.1
Arriba a la izquierda = más inteligencia por dólar. Eje X logarítmico.
| # | Modelo | Intelligence | Coding | Agentic | Contexto | USD/1M (in/out) | Int./USD |
|---|---|---|---|---|---|---|---|
| ★ | Anthropic: Claude Fable 5.1 anthropic/claude-fable-5.1 | 53.4 | 81.6 | 58.0 | 1.0M | $10.00 / $50.00 batch: $5.00 / $25.00 | 1.1 |
| 02 | OpenAI: GPT-6 Astra openai/gpt-6-astra | 52.8 | 76.9 | 51.5 | 1.1M | $10.00 / $50.00 batch: $5.00 / $25.00 | 1.1 |
| 03 | Claude Opus 5 anthropic/claude-opus-5 | 50.7 | 78.0 | 56.2 | 1.0M | $5.00 / $25.00 batch: $2.50 / $12.50 | 2.0 |
| 04 | Anthropic: Claude Fable 5 anthropic/claude-fable-5 | 49.7 | 76.5 | 51.0 | 1.0M | $10.00 / $50.00 batch: $5.00 / $25.00 | 1.0 |
| 05 | OpenAI: GPT-5.6 Sol openai/gpt-5.6-sol | 47.1 | 77.4 | 50.5 | 1.1M | $2.00 / $10.00 batch: $1.00 / $5.00 | 4.7 |
| 06 | Z.ai: GLM 5.3 z-ai/glm-5.3 | 44.9 | 74.8 | 53.4 | 1.3M | $1.40 / $4.40 batch: $0.70 / $2.20 | 10.2 |
| 07 | SpaceXAI: Grok 4.6 x-ai/grok-4.6 | 44.4 | 76.8 | 53.4 | 500K | $2.00 / $6.00 | 7.4 |
| 08 | MoonshotAI: Kimi K3 moonshotai/kimi-k3 | 43.8 | 76.2 | 50.6 | 1.0M | $2.34 / $11.70 batch: $3.00 / $15.00 | 3.7 |
| 09 | OpenAI: GPT-5.6 Terra openai/gpt-5.6-terra | 42.3 | 76.7 | 43.7 | 1.1M | $2.00 / $12.00 batch: $1.00 / $6.00 | 3.5 |
| 10 | Anthropic: Claude Opus 4.8 anthropic/claude-opus-4.8 | 42.0 | 74.3 | 42.6 | 1.0M | $5.00 / $25.00 batch: $2.50 / $12.50 | 1.7 |
| 11 | Z.ai: GLM 5.3 Flash z-ai/glm-5.3-flash | 41.9 | 71.5 | 51.2 | 1.3M | $0.15 / $0.50 batch: $0.07 / $0.25 | 83.8 |
| 12 | Google: Gemini 3.8 Flash google/gemini-3.8-flash | 41.2 | 76.3 | 41.1 | 1.0M | $0.75 / $3.75 batch: $0.38 / $1.88 | 11.0 |
| 13 | Qwen: Qwen3.8 Max (0902) qwen/qwen3.8-max-0902 | 40.3 | 71.8 | 49.6 | 1.0M | $2.00 / $6.00 | 6.7 |
| 14 | Qwen: Qwen3.8 2.4T A95B qwen/qwen3.8-2.4t-a95b | 40.0 | 71.9 | 50.4 | 1.0M | $2.00 / $6.00 batch: $2.00 / $6.00 | 6.7 |
| 15 | Meta: Muse Spark 1.2 meta/muse-spark-1.2 | 39.8 | 72.2 | 44.0 | 1.0M | $1.25 / $4.25 | 9.4 |
| 16 | DeepSeek: DeepSeek V4.1 Flash deepseek/deepseek-v4.1-flash | 39.5 | — | — | 1.0M | $0.15 / $0.60 | 65.8 |
| 17 | Google: Gemini 3.7 Flash google/gemini-3.7-flash | 39.4 | 76.1 | 36.4 | 1.0M | $0.75 / $3.75 batch: $0.38 / $1.88 | 10.5 |
| 18 | SpaceXAI: Grok 4.5 x-ai/grok-4.5 | 39.1 | 72.4 | 42.1 | 500K | $2.00 / $6.00 | 6.5 |
| 19 | OpenAI: GPT-5.5 openai/gpt-5.5 | 38.6 | 74.9 | 37.3 | 1.1M | $5.00 / $30.00 batch: $2.50 / $15.00 | 1.3 |
| 20 | Anthropic: Claude Sonnet 5 anthropic/claude-sonnet-5 | 38.4 | 71.5 | 44.3 | 1.0M | $2.00 / $10.00 batch: $1.00 / $5.00 | 3.8 |
Cómo se construye este radar
Los datos provienen de la API pública de OpenRouter,
que expone los índices de Artificial Analysis.
El JSON se regenera cada semana con un script versionado y validado
(scripts/refresh-radar.mjs).