Artículo
Amazon Bedrock: Inteligencia vs Costo
El catálogo de Amazon Bedrock no para de expandirse: hoy reúne modelos de más de 15 proveedores líderes de inteligencia artificial —desde gigantes como Anthropic, OpenAI, Meta, Google y Amazon, hasta referentes open source como DeepSeek, Mistral, Qwen y Z.AI—. Sin embargo, para cualquier equipo de ingeniería o producto el dilema sigue siendo el mismo: ¿cómo elegir el modelo más costo-eficiente y saber en qué casos realmente vale la pena pagar más?
Para responder a esto, este análisis combina los precios oficiales on-demand de Bedrock[1] con las métricas públicas del Intelligence Index de Artificial Analysis (v4.3.2)[2]. El estudio abarca 68 modelos únicos evaluados en 146 configuraciones (desglosando cada nivel de esfuerzo de razonamiento disponible), con fecha de corte al 24 de septiembre de 2026. Incluye las incorporaciones más recientes como Claude Opus 5.5 y la nueva generación de OpenAI con GPT-6 Luna y GPT-6 Sol.
Cómo leer el gráfico
El gráfico enfrenta el costo de entrada frente a la capacidad de cada modelo:
- Eje X (Costo): Precio por millón de tokens de entrada en escala logarítmica (USD, lista on-demand de Bedrock).
- Eje Y (Inteligencia): Puntuación en el Intelligence Index.
- Puntos y variantes: Cada punto representa un modelo. Cuando un modelo ofrece varios niveles de razonamiento (low, medium, high, max), se grafica un punto por cada nivel en la misma coordenada X: en Bedrock el costo por token de entrada no varía según el nivel de razonamiento, por lo que ese incremento vertical representa inteligencia adicional sin sobrecosto en el precio base de entrada.
- Frontera de Pareto: Al activar el interruptor, se traza la línea de eficiencia óptima. Conecta los modelos para los cuales no existe alternativa en el catálogo que sea simultáneamente más barata y más inteligente.
Usa la rueda o pellizca para hacer zoom · arrastra para explorar · doble clic para restablecer
Precio on-demand de entrada y salida ($/1M tokens) extraído de Amazon Bedrock us-east-1. Índice de inteligencia: Artificial Analysis Intelligence Index v4.3.2. Puedes alternar la vista entre precios de Tokens de Entrada y Tokens de Salida usando el selector superior. Un punto sobre la frontera no es dominado en precio ni en índice por ningún otro punto del set. Datos con corte al 24 de septiembre de 2026.
Top 10 por Intelligence Index
- 1Claude Opus 5.5 · maxIn: $4 · Out: $2057.6
- 2Claude Opus 5.5 · xhighIn: $4 · Out: $2056.0
- 3Claude Opus 5.5 · highIn: $4 · Out: $2053.6
- 4Claude Fable 5.1 · maxIn: $11 · Out: $5553.4
- 5Claude Fable 5.1 · xhighIn: $11 · Out: $5553.2
- 6GPT-6 Astra · maxIn: $11 · Out: $5552.7
- 7GPT-6 Astra · xhighIn: $11 · Out: $5552.4
- 8Claude Opus 5.5 · mediumIn: $4 · Out: $2051.2
- 9Claude Fable 5.1 · highIn: $11 · Out: $5551.1
- 10GPT-6 Astra · highIn: $11 · Out: $5550.9
Ver los 146 puntos en tabla
| Modelo | Esfuerzo | Proveedor | $/1M in | $/1M out | Índice |
|---|---|---|---|---|---|
| Nova Micro | — | Amazon | $0.035 | $0.14 | 5.88 |
| Gemma 3 4B IT | — | $0.04 | $0.08 | 4.83 | |
| Nemotron Nano 3 30B | reasoning | NVIDIA | $0.06 | $0.24 | 8.90 |
| NVIDIA Nemotron Nano 9B v2 | reasoning | NVIDIA | $0.06 | $0.23 | 7.43 |
| Nemotron Nano 3 30B | non-reasoning | NVIDIA | $0.06 | $0.24 | 6.85 |
| NVIDIA Nemotron Nano 9B v2 | non-reasoning | NVIDIA | $0.06 | $0.23 | 6.84 |
| Nova Lite | — | Amazon | $0.06 | $0.24 | 6.67 |
| GLM 4.7 Flash | reasoning | Z.AI | $0.07 | $0.4 | 14.87 |
| GLM 4.7 Flash | non-reasoning | Z.AI | $0.07 | $0.4 | 10.56 |
| gpt-oss-20b | low | OpenAI | $0.07 | $0.3 | 9.95 |
| gpt-oss-20b | high | OpenAI | $0.07 | $0.3 | 8.97 |
| Gemma 3 12B IT | — | $0.09 | $0.29 | 3.75 | |
| Ministral 3B | — | Mistral AI | $0.1 | $0.1 | 4.84 |
| GPT-6 Luna | max | OpenAI | $0.11 | $0.55 | 37.26 |
| GPT-6 Luna | xhigh | OpenAI | $0.11 | $0.55 | 33.88 |
| GPT-6 Luna | high | OpenAI | $0.11 | $0.55 | 32.15 |
| GPT-6 Luna | medium | OpenAI | $0.11 | $0.55 | 29.46 |
| GPT-6 Luna | low | OpenAI | $0.11 | $0.55 | 20.92 |
| GPT-6 Luna | non-reasoning | OpenAI | $0.11 | $0.55 | 18.26 |
| NVIDIA Nemotron 3 Super 120B A12B | reasoning | NVIDIA | $0.15 | $0.65 | 12.83 |
| gpt-oss-120b | high | OpenAI | $0.15 | $0.6 | 11.60 |
| Qwen3 Next 80B A3B | reasoning | Qwen | $0.15 | $1.2 | 11.20 |
| gpt-oss-120b | low | OpenAI | $0.15 | $0.6 | 10.21 |
| Qwen3-Coder-30B-A3B-Instruct | — | Qwen | $0.15 | $0.6 | 9.59 |
| Qwen3 32B (dense) | reasoning | Qwen | $0.15 | $0.6 | 8.59 |
| Qwen3 32B (dense) | non-reasoning | Qwen | $0.15 | $0.6 | 7.34 |
| Ministral 3 8B | — | Mistral AI | $0.15 | $0.15 | 5.48 |
| Mistral 7B Instruct | — | Mistral AI | $0.15 | $0.2 | 5.03 |
| Llama 4 Scout 17B Instruct | — | Meta | $0.17 | $0.66 | 8.08 |
| NVIDIA Nemotron Nano 12B v2 VL BF16 | reasoning | NVIDIA | $0.2 | $0.6 | 7.48 |
| Ministral 14B 3.0 | — | Mistral AI | $0.2 | $0.2 | 6.04 |
| NVIDIA Nemotron Nano 12B v2 VL BF16 | non-reasoning | NVIDIA | $0.2 | $0.6 | 5.82 |
| GPT-5.6 Luna | max | OpenAI | $0.22 | $1.32 | 37.32 |
| GPT-5.6 Luna | xhigh | OpenAI | $0.22 | $1.32 | 34.56 |
| GPT-5.6 Luna | high | OpenAI | $0.22 | $1.32 | 32.12 |
| GPT-5.6 Luna | medium | OpenAI | $0.22 | $1.32 | 25.04 |
| GPT-5.6 Luna | low | OpenAI | $0.22 | $1.32 | 21.01 |
| GPT-5.6 Luna | non-reasoning | OpenAI | $0.22 | $1.32 | 15.53 |
| Llama 3.1 8B Instruct | — | Meta | $0.22 | $0.22 | 6.93 |
| Gemma 3 27B PT | — | $0.23 | $0.38 | 4.85 | |
| Llama 4 Maverick 17B Instruct | — | Meta | $0.24 | $0.97 | 9.99 |
| MiniMax M2.5 | — | MiniMax | $0.3 | $1.2 | 22.80 |
| MiniMax M2.1 | — | MiniMax | $0.3 | $1.2 | 20.95 |
| MiniMax M2 | — | MiniMax | $0.3 | $1.2 | 18.63 |
| Nova 2 Lite | high | Amazon | $0.3 | $2.5 | 13.37 |
| Nova 2 Lite | medium | Amazon | $0.3 | $2.5 | 12.48 |
| Nova 2 Lite | low | Amazon | $0.3 | $2.5 | 11.80 |
| Nova 2 Lite | non-reasoning | Amazon | $0.3 | $2.5 | 8.74 |
| Llama 3 8B Instruct | — | Meta | $0.3 | $0.6 | 4.83 |
| Devstral 2 123B | — | Mistral AI | $0.4 | $2 | 8.62 |
| Mixtral 8x7B Instruct | — | Mistral AI | $0.45 | $0.7 | 5.12 |
| Mistral Large 3 | — | Mistral AI | $0.5 | $1.5 | 9.27 |
| Qwen3 Coder Next | — | Qwen | $0.5 | $1.2 | 9.24 |
| Magistral Small 2509 | — | Mistral AI | $0.5 | $1.5 | 8.59 |
| Qwen3 VL 235B A22B | reasoning | Qwen | $0.53 | $2.66 | 13.44 |
| Kimi K2.5 | reasoning | Moonshot AI | $0.6 | $3 | 23.46 |
| GLM 4.7 | reasoning | Z.AI | $0.6 | $2.2 | 22.24 |
| Kimi K2 Thinking | — | Moonshot AI | $0.6 | $2.5 | 22.02 |
| Kimi K2.5 | non-reasoning | Moonshot AI | $0.6 | $3 | 19.43 |
| GLM 4.7 | non-reasoning | Z.AI | $0.6 | $2.2 | 17.36 |
| DeepSeek V3.2 | reasoning | DeepSeek | $0.62 | $1.85 | 21.49 |
| DeepSeek V3.2 | non-reasoning | DeepSeek | $0.62 | $1.85 | 16.04 |
| Llama 3.3 70B Instruct | — | Meta | $0.72 | $0.72 | 7.66 |
| Llama 3.1 70B Instruct | — | Meta | $0.72 | $0.72 | 6.60 |
| Nova Pro | — | Amazon | $0.8 | $3.2 | 6.96 |
| GLM 5 | reasoning | Z.AI | $1 | $3.2 | 27.91 |
| GLM 5 | non-reasoning | Z.AI | $1 | $3.2 | 21.78 |
| Claude Haiku 4.5 | reasoning | Anthropic | $1 | $5 | 16.88 |
| Claude Haiku 4.5 | non-reasoning | Anthropic | $1 | $5 | 15.41 |
| Mistral Small (24.02) | Sep '24 | Mistral AI | $1 | $3 | 5.85 |
| Mistral Small (24.02) | Feb '24 | Mistral AI | $1 | $3 | 5.50 |
| DeepSeek-R1 | Jan '25 | DeepSeek | $1.35 | $5.4 | 11.41 |
| Pixtral Large (25.02) | — | Mistral AI | $2 | $6 | 7.15 |
| GPT-6 Sol | max | OpenAI | $2.2 | $11 | 47.53 |
| Grok 4.6 | high | xAI | $2.2 | $6.6 | 44.31 |
| Grok 4.6 | xhigh | xAI | $2.2 | $6.6 | 44.20 |
| GPT-6 Sol | xhigh | OpenAI | $2.2 | $11 | 44.10 |
| Grok 4.6 | medium | xAI | $2.2 | $6.6 | 42.84 |
| GPT-6 Sol | high | OpenAI | $2.2 | $11 | 42.82 |
| GPT-5.6 Terra | max | OpenAI | $2.2 | $13.2 | 42.08 |
| GPT-6 Sol | medium | OpenAI | $2.2 | $11 | 39.78 |
| Claude Sonnet 5 | max | Anthropic | $2.2 | $11 | 38.16 |
| GPT-5.6 Terra | xhigh | OpenAI | $2.2 | $13.2 | 37.95 |
| Grok 4.6 | low | xAI | $2.2 | $6.6 | 35.12 |
| Claude Sonnet 5 | xhigh | Anthropic | $2.2 | $11 | 34.38 |
| GPT-5.6 Terra | high | OpenAI | $2.2 | $13.2 | 34.24 |
| GPT-6 Sol | low | OpenAI | $2.2 | $11 | 33.90 |
| Claude Sonnet 5 | high | Anthropic | $2.2 | $11 | 31.66 |
| GPT-5.6 Terra | medium | OpenAI | $2.2 | $13.2 | 30.09 |
| GPT-6 Sol | non-reasoning | OpenAI | $2.2 | $11 | 28.09 |
| Claude Sonnet 5 | medium | Anthropic | $2.2 | $11 | 28.05 |
| GPT-5.6 Terra | low | OpenAI | $2.2 | $13.2 | 27.50 |
| Claude Sonnet 5 | low | Anthropic | $2.2 | $11 | 24.26 |
| Claude Sonnet 5 | high | Anthropic | $2.2 | $11 | 23.20 |
| GPT-5.6 Terra | non-reasoning | OpenAI | $2.2 | $13.2 | 20.78 |
| Llama 3 70B Instruct | — | Meta | $2.65 | $3.5 | 5.45 |
| GPT 5.4 | xhigh | OpenAI | $2.75 | $16.5 | 38.98 |
| GPT 5.4 | low | OpenAI | $2.75 | $16.5 | 27.58 |
| GPT 5.4 | non-reasoning | OpenAI | $2.75 | $16.5 | 18.16 |
| Claude Sonnet 4.6 | max | Anthropic | $3 | $15 | 30.06 |
| Claude Sonnet 4.6 | high | Anthropic | $3 | $15 | 24.69 |
| Claude Sonnet 4.6 | low | Anthropic | $3 | $15 | 23.30 |
| Claude Sonnet 4.5 | reasoning | Anthropic | $3 | $15 | 20.67 |
| Claude Sonnet 4.5 | non-reasoning | Anthropic | $3 | $15 | 19.34 |
| Kimi K3 | max | Moonshot AI | $3.3 | $16.5 | 43.59 |
| Kimi K3 | low | Moonshot AI | $3.3 | $16.5 | 34.46 |
| Claude Opus 5.5 | max | Anthropic | $4 | $20 | 57.62 |
| Claude Opus 5.5 | xhigh | Anthropic | $4 | $20 | 55.99 |
| Claude Opus 5.5 | high | Anthropic | $4 | $20 | 53.58 |
| Claude Opus 5.5 | medium | Anthropic | $4 | $20 | 51.24 |
| Claude Opus 5.5 | low | Anthropic | $4 | $20 | 42.31 |
| Mistral Large (24.02) | Feb '24 | Mistral AI | $4 | $12 | 5.76 |
| GPT-5.6 Sol | max | OpenAI | $4.4 | $22 | 46.97 |
| GPT-5.6 Sol | xhigh | OpenAI | $4.4 | $22 | 44.01 |
| GPT-5.6 Sol | high | OpenAI | $4.4 | $22 | 42.35 |
| GPT-5.6 Sol | medium | OpenAI | $4.4 | $22 | 39.24 |
| GPT-5.6 Sol | low | OpenAI | $4.4 | $22 | 33.47 |
| GPT-5.6 Sol | non-reasoning | OpenAI | $4.4 | $22 | 28.33 |
| Claude Opus 4.6 | max | Anthropic | $5 | $25 | 31.95 |
| Claude Opus 4.5 | reasoning | Anthropic | $5 | $25 | 29.10 |
| Claude Opus 4.6 | high | Anthropic | $5 | $25 | 26.35 |
| Claude Opus 4.5 | non-reasoning | Anthropic | $5 | $25 | 23.68 |
| Claude Opus 5 | max | Anthropic | $5.5 | $27.5 | 50.78 |
| Claude Opus 5 | xhigh | Anthropic | $5.5 | $27.5 | 49.68 |
| Claude Opus 5 | high | Anthropic | $5.5 | $27.5 | 48.12 |
| Claude Opus 5 | medium | Anthropic | $5.5 | $27.5 | 44.83 |
| Claude Opus 4.8 | max | Anthropic | $5.5 | $27.5 | 41.79 |
| Claude Opus 4.7 | max | Anthropic | $5.5 | $27.5 | 40.69 |
| Claude Opus 5 | low | Anthropic | $5.5 | $27.5 | 39.35 |
| GPT 5.5 | xhigh | OpenAI | $5.5 | $33 | 38.36 |
| GPT 5.5 | high | OpenAI | $5.5 | $33 | 36.98 |
| GPT 5.5 | medium | OpenAI | $5.5 | $33 | 33.80 |
| Claude Opus 4.7 | high | Anthropic | $5.5 | $27.5 | 30.93 |
| GPT 5.5 | low | OpenAI | $5.5 | $33 | 30.71 |
| GPT 5.5 | non-reasoning | OpenAI | $5.5 | $33 | 23.17 |
| Claude Fable 5.1 | max | Anthropic | $11 | $55 | 53.35 |
| Claude Fable 5.1 | xhigh | Anthropic | $11 | $55 | 53.20 |
| GPT-6 Astra | max | OpenAI | $11 | $55 | 52.67 |
| GPT-6 Astra | xhigh | OpenAI | $11 | $55 | 52.39 |
| Claude Fable 5.1 | high | Anthropic | $11 | $55 | 51.15 |
| GPT-6 Astra | high | OpenAI | $11 | $55 | 50.92 |
| Claude Fable 5 | max | Anthropic | $11 | $55 | 49.63 |
| GPT-6 Astra | medium | OpenAI | $11 | $55 | 49.57 |
| Claude Fable 5.1 | medium | Anthropic | $11 | $55 | 48.92 |
| Claude Fable 5.1 | low | Anthropic | $11 | $55 | 46.82 |
| GPT-6 Astra | low | OpenAI | $11 | $55 | 45.78 |
Hallazgos clave en la frontera
De las 146 configuraciones analizadas, solo 7 se sitúan sobre la frontera de Pareto. De este grupo se desprenden conclusiones contundentes tras la llegada de Claude Opus 5.5, GPT-6 Luna y GPT-6 Sol:
- Claude Opus 5.5 redefine el techo de la frontera: La incorporación de Claude Opus 5.5 a $4.00 por millón de tokens de entrada ($20.00/1M salida) y 1M de ventana de contexto altera drásticamente la relación costo-beneficio en la gama alta. Con un índice de 57.62 en razonamiento máximo (max), no solo se consolida como el modelo más potente del catálogo, sino que lo hace costando menos que Claude Opus 5 ($5.50), y a un tercio del precio de alternativas como Claude Fable 5.1 ($11.00, índice 53.35) o GPT-6 Astra ($11.00, índice 52.67). Su eficiencia dominante expulsa a los modelos previos de gama ultra-alta de la frontera de Pareto.
- GPT-6 Luna y Sol conquistan la frontera: OpenAI introduce GPT-6 Luna a tan solo $0.11/1M tokens alcanzando un índice de 37.26 en modo max, lo que supone prácticamente igualar la inteligencia de GPT-5.6 Luna a la mitad de su costo. Por su parte, GPT-6 Sol entra a $2.20/1M con un índice de 47.53 en max, superando directamente a Grok 4.6 (44.31) y Claude Sonnet 5 (38.16) a ese mismo precio y apoderándose del escalón intermedio-alto de la frontera.
- El esfuerzo de razonamiento aporta más que migrar de modelo: Tanto en GPT-6 Luna como en GPT-5.6 Luna, activar el esfuerzo de razonamiento eleva el índice de 18.3 a 37.3 sin costo adicional por token de entrada base. Esta ganancia vertical supera la brecha entre modelos enteros de mayor precio sin razonamiento.
- Por debajo de $0.10/1M, dominan los modelos compactos y ultra-ligeros: Amazon Nova Micro ($0.035, índice 5.88), NVIDIA Nemotron Nano 3 30B con razonamiento ($0.06, índice 8.90) y GLM 4.7 Flash ($0.07, índice 14.87) dominan con autoridad el segmento de ultra bajo costo; nadie ofrece más inteligencia por menos dinero en esa franja antes del salto a GPT-6 Luna.
Precio por inteligencia: ¿cuánto cuesta realmente cada punto de benchmark?
Para profundizar en la economía de Bedrock más allá de la frontera de Pareto, evaluamos directamente el costo por punto de inteligencia ($ / punto). Esta métrica divide el precio por millón de tokens (ya sea de entrada o de salida) entre la puntuación del modelo en el Intelligence Index:
¿Cómo interpretarlo? Mide cuánto pagas en tu factura por millón de tokens por cada punto de benchmark que el modelo te ofrece. Menor es mejor: un menor costo por punto indica que el modelo es mucho más eficiente económicamente (compras más inteligencia por cada dólar invertido). Un costo elevado revela modelos antiguos con sobreprecio desproporcionado.
Tomando como referencia los tokens de entrada (y recordando que puedes alternar a tokens de salida en el comparador interactivo inferior), los contrastes son contundentes:
- GPT-6 Luna [max] es el modelo más eficiente por punto de todo Bedrock: Con un índice de 37.26 por solo $0.11/1M tokens, cada punto de benchmark en GPT-6 Luna (max) cuesta apenas $0.0030 por punto. Supera a GLM 4.7 Flash ($0.0047/pt) y pulveriza el costo promedio del catálogo ($0.090/pt) por un factor de 30x.
- La brecha entre tiers: En el segmento de entrada (< $0.50/1M), el costo promedio es de solo $0.021 por punto. En contraste, en el segmento insignia o flagship (> $2.50/1M), el costo promedio asciende a $0.180 por punto. Comprar inteligencia en la gama alta cuesta, en promedio, casi nueve veces más por cada punto de capacidad adicional.
- Las trampas de sobreprecio (modelos desactualizados): Opciones de generaciones pasadas como Mistral Large 24.02 ($4.00/1M, índice 5.76) y Llama 3 70B ($2.65/1M, índice 5.45) exhiben costos desorbitados de $0.694 por punto y $0.486 por punto, respectivamente. Tienen un sobreprecio de más de 160 a 230 veces frente a GPT-6 Luna, debido a que sus tarifas de lista nunca fueron reducidas tras los lanzamientos posteriores.
- La anomalía económica de Claude Opus 5.5: En el segmento flagship, donde la media de costo ronda los $0.180/pt, el nuevo Claude Opus 5.5 entra a solo $0.0694 por punto con un puntaje récord de 57.62. Es casi tres veces más eficiente por punto que alternativas de su misma franja como Claude Fable 5.1 ($0.206/pt).
Costo por Punto de Inteligencia
Compara la relación calidad-precio en cada tier. Menor costo por punto equivale a mayor eficiencia.
- 1GPT-6 Luna· max$0.11/1M · 37.3 pts$0.0030/ punto
- 2GPT-6 Luna· xhigh$0.11/1M · 33.9 pts$0.0032/ punto
- 3GPT-6 Luna· high$0.11/1M · 32.1 pts$0.0034/ punto
- 4GPT-6 Luna· medium$0.11/1M · 29.5 pts$0.0037/ punto
- 5GLM 4.7 Flash· reasoning$0.07/1M · 14.9 pts$0.0047/ punto
- 6GPT-6 Luna· low$0.11/1M · 20.9 pts$0.0053/ punto
- 1Mistral Large (24.02)· Feb '24$4/1M · 5.8 pts$0.6944/ punto
- 2Llama 3 70B Instruct$2.65/1M · 5.5 pts$0.4862/ punto
- 3Pixtral Large (25.02)$2/1M · 7.2 pts$0.2797/ punto
- 4GPT-6 Astra· low$11/1M · 45.8 pts$0.2403/ punto
- 5GPT 5.5· non-reasoning$5.5/1M · 23.2 pts$0.2374/ punto
- 6Claude Fable 5.1· low$11/1M · 46.8 pts$0.2349/ punto
Ver modelos de este tier en tabla (146)
| Modelo | Proveedor | $/1M In | Índice | Costo/Pt ↑ |
|---|---|---|---|---|
| GPT-6 Luna(max) | OpenAI | $0.11 | 37.26 | $0.0030 |
| GPT-6 Luna(xhigh) | OpenAI | $0.11 | 33.88 | $0.0032 |
| GPT-6 Luna(high) | OpenAI | $0.11 | 32.15 | $0.0034 |
| GPT-6 Luna(medium) | OpenAI | $0.11 | 29.46 | $0.0037 |
| GLM 4.7 Flash(reasoning) | Z.AI | $0.07 | 14.87 | $0.0047 |
| GPT-6 Luna(low) | OpenAI | $0.11 | 20.92 | $0.0053 |
| GPT-5.6 Luna(max) | OpenAI | $0.22 | 37.32 | $0.0059 |
| Nova Micro | Amazon | $0.035 | 5.88 | $0.0060 |
| GPT-6 Luna(non-reasoning) | OpenAI | $0.11 | 18.26 | $0.0060 |
| GPT-5.6 Luna(xhigh) | OpenAI | $0.22 | 34.56 | $0.0064 |
| GLM 4.7 Flash(non-reasoning) | Z.AI | $0.07 | 10.56 | $0.0066 |
| Nemotron Nano 3 30B(reasoning) | NVIDIA | $0.06 | 8.90 | $0.0067 |
| GPT-5.6 Luna(high) | OpenAI | $0.22 | 32.12 | $0.0068 |
| gpt-oss-20b(low) | OpenAI | $0.07 | 9.95 | $0.0070 |
| gpt-oss-20b(high) | OpenAI | $0.07 | 8.97 | $0.0078 |
| NVIDIA Nemotron Nano 9B v2(reasoning) | NVIDIA | $0.06 | 7.43 | $0.0081 |
| Gemma 3 4B IT | $0.04 | 4.83 | $0.0083 | |
| Nemotron Nano 3 30B(non-reasoning) | NVIDIA | $0.06 | 6.85 | $0.0088 |
| NVIDIA Nemotron Nano 9B v2(non-reasoning) | NVIDIA | $0.06 | 6.84 | $0.0088 |
| GPT-5.6 Luna(medium) | OpenAI | $0.22 | 25.04 | $0.0088 |
| Nova Lite | Amazon | $0.06 | 6.67 | $0.0090 |
| GPT-5.6 Luna(low) | OpenAI | $0.22 | 21.01 | $0.0105 |
| NVIDIA Nemotron 3 Super 120B A12B(reasoning) | NVIDIA | $0.15 | 12.83 | $0.0117 |
| gpt-oss-120b(high) | OpenAI | $0.15 | 11.60 | $0.0129 |
| MiniMax M2.5 | MiniMax | $0.3 | 22.80 | $0.0132 |
| Qwen3 Next 80B A3B(reasoning) | Qwen | $0.15 | 11.20 | $0.0134 |
| GPT-5.6 Luna(non-reasoning) | OpenAI | $0.22 | 15.53 | $0.0142 |
| MiniMax M2.1 | MiniMax | $0.3 | 20.95 | $0.0143 |
| gpt-oss-120b(low) | OpenAI | $0.15 | 10.21 | $0.0147 |
| Qwen3-Coder-30B-A3B-Instruct | Qwen | $0.15 | 9.59 | $0.0156 |
| MiniMax M2 | MiniMax | $0.3 | 18.63 | $0.0161 |
| Qwen3 32B (dense)(reasoning) | Qwen | $0.15 | 8.59 | $0.0175 |
| Qwen3 32B (dense)(non-reasoning) | Qwen | $0.15 | 7.34 | $0.0204 |
| Ministral 3B | Mistral AI | $0.1 | 4.84 | $0.0207 |
| Llama 4 Scout 17B Instruct | Meta | $0.17 | 8.08 | $0.0210 |
| Nova 2 Lite(high) | Amazon | $0.3 | 13.37 | $0.0224 |
| Gemma 3 12B IT | $0.09 | 3.75 | $0.0240 | |
| Llama 4 Maverick 17B Instruct | Meta | $0.24 | 9.99 | $0.0240 |
| Nova 2 Lite(medium) | Amazon | $0.3 | 12.48 | $0.0240 |
| Nova 2 Lite(low) | Amazon | $0.3 | 11.80 | $0.0254 |
| Kimi K2.5(reasoning) | Moonshot AI | $0.6 | 23.46 | $0.0256 |
| NVIDIA Nemotron Nano 12B v2 VL BF16(reasoning) | NVIDIA | $0.2 | 7.48 | $0.0267 |
| GLM 4.7(reasoning) | Z.AI | $0.6 | 22.24 | $0.0270 |
| Kimi K2 Thinking | Moonshot AI | $0.6 | 22.02 | $0.0272 |
| Ministral 3 8B | Mistral AI | $0.15 | 5.48 | $0.0274 |
| DeepSeek V3.2(reasoning) | DeepSeek | $0.62 | 21.49 | $0.0289 |
| Mistral 7B Instruct | Mistral AI | $0.15 | 5.03 | $0.0298 |
| Kimi K2.5(non-reasoning) | Moonshot AI | $0.6 | 19.43 | $0.0309 |
| Llama 3.1 8B Instruct | Meta | $0.22 | 6.93 | $0.0317 |
| Ministral 14B 3.0 | Mistral AI | $0.2 | 6.04 | $0.0331 |
| Nova 2 Lite(non-reasoning) | Amazon | $0.3 | 8.74 | $0.0343 |
| NVIDIA Nemotron Nano 12B v2 VL BF16(non-reasoning) | NVIDIA | $0.2 | 5.82 | $0.0344 |
| GLM 4.7(non-reasoning) | Z.AI | $0.6 | 17.36 | $0.0346 |
| GLM 5(reasoning) | Z.AI | $1 | 27.91 | $0.0358 |
| DeepSeek V3.2(non-reasoning) | DeepSeek | $0.62 | 16.04 | $0.0387 |
| Qwen3 VL 235B A22B(reasoning) | Qwen | $0.53 | 13.44 | $0.0394 |
| GLM 5(non-reasoning) | Z.AI | $1 | 21.78 | $0.0459 |
| GPT-6 Sol(max) | OpenAI | $2.2 | 47.53 | $0.0463 |
| Devstral 2 123B | Mistral AI | $0.4 | 8.62 | $0.0464 |
| Gemma 3 27B PT | $0.23 | 4.85 | $0.0474 | |
| Grok 4.6(high) | xAI | $2.2 | 44.31 | $0.0497 |
| Grok 4.6(xhigh) | xAI | $2.2 | 44.20 | $0.0498 |
| GPT-6 Sol(xhigh) | OpenAI | $2.2 | 44.10 | $0.0499 |
| Grok 4.6(medium) | xAI | $2.2 | 42.84 | $0.0514 |
| GPT-6 Sol(high) | OpenAI | $2.2 | 42.82 | $0.0514 |
| GPT-5.6 Terra(max) | OpenAI | $2.2 | 42.08 | $0.0523 |
| Mistral Large 3 | Mistral AI | $0.5 | 9.27 | $0.0539 |
| Qwen3 Coder Next | Qwen | $0.5 | 9.24 | $0.0541 |
| GPT-6 Sol(medium) | OpenAI | $2.2 | 39.78 | $0.0553 |
| Claude Sonnet 5(max) | Anthropic | $2.2 | 38.16 | $0.0577 |
| GPT-5.6 Terra(xhigh) | OpenAI | $2.2 | 37.95 | $0.0580 |
| Magistral Small 2509 | Mistral AI | $0.5 | 8.59 | $0.0582 |
| Claude Haiku 4.5(reasoning) | Anthropic | $1 | 16.88 | $0.0592 |
| Llama 3 8B Instruct | Meta | $0.3 | 4.83 | $0.0621 |
| Grok 4.6(low) | xAI | $2.2 | 35.12 | $0.0626 |
| Claude Sonnet 5(xhigh) | Anthropic | $2.2 | 34.38 | $0.0640 |
| GPT-5.6 Terra(high) | OpenAI | $2.2 | 34.24 | $0.0643 |
| Claude Haiku 4.5(non-reasoning) | Anthropic | $1 | 15.41 | $0.0649 |
| GPT-6 Sol(low) | OpenAI | $2.2 | 33.90 | $0.0649 |
| Claude Opus 5.5(max) | Anthropic | $4 | 57.62 | $0.0694 |
| Claude Sonnet 5(high) | Anthropic | $2.2 | 31.66 | $0.0695 |
| GPT 5.4(xhigh) | OpenAI | $2.75 | 38.98 | $0.0705 |
| Claude Opus 5.5(xhigh) | Anthropic | $4 | 55.99 | $0.0714 |
| GPT-5.6 Terra(medium) | OpenAI | $2.2 | 30.09 | $0.0731 |
| Claude Opus 5.5(high) | Anthropic | $4 | 53.58 | $0.0747 |
| Kimi K3(max) | Moonshot AI | $3.3 | 43.59 | $0.0757 |
| Claude Opus 5.5(medium) | Anthropic | $4 | 51.24 | $0.0781 |
| GPT-6 Sol(non-reasoning) | OpenAI | $2.2 | 28.09 | $0.0783 |
| Claude Sonnet 5(medium) | Anthropic | $2.2 | 28.05 | $0.0784 |
| GPT-5.6 Terra(low) | OpenAI | $2.2 | 27.50 | $0.0800 |
| Mixtral 8x7B Instruct | Mistral AI | $0.45 | 5.12 | $0.0879 |
| Claude Sonnet 5(low) | Anthropic | $2.2 | 24.26 | $0.0907 |
| GPT-5.6 Sol(max) | OpenAI | $4.4 | 46.97 | $0.0937 |
| Llama 3.3 70B Instruct | Meta | $0.72 | 7.66 | $0.0940 |
| Claude Opus 5.5(low) | Anthropic | $4 | 42.31 | $0.0945 |
| Claude Sonnet 5(high) | Anthropic | $2.2 | 23.20 | $0.0948 |
| Kimi K3(low) | Moonshot AI | $3.3 | 34.46 | $0.0958 |
| GPT 5.4(low) | OpenAI | $2.75 | 27.58 | $0.0997 |
| Claude Sonnet 4.6(max) | Anthropic | $3 | 30.06 | $0.0998 |
| GPT-5.6 Sol(xhigh) | OpenAI | $4.4 | 44.01 | $0.1000 |
| GPT-5.6 Sol(high) | OpenAI | $4.4 | 42.35 | $0.1039 |
| GPT-5.6 Terra(non-reasoning) | OpenAI | $2.2 | 20.78 | $0.1059 |
| Claude Opus 5(max) | Anthropic | $5.5 | 50.78 | $0.1083 |
| Llama 3.1 70B Instruct | Meta | $0.72 | 6.60 | $0.1091 |
| Claude Opus 5(xhigh) | Anthropic | $5.5 | 49.68 | $0.1107 |
| GPT-5.6 Sol(medium) | OpenAI | $4.4 | 39.24 | $0.1121 |
| Claude Opus 5(high) | Anthropic | $5.5 | 48.12 | $0.1143 |
| Nova Pro | Amazon | $0.8 | 6.96 | $0.1149 |
| DeepSeek-R1(Jan '25) | DeepSeek | $1.35 | 11.41 | $0.1183 |
| Claude Sonnet 4.6(high) | Anthropic | $3 | 24.69 | $0.1215 |
| Claude Opus 5(medium) | Anthropic | $5.5 | 44.83 | $0.1227 |
| Claude Sonnet 4.6(low) | Anthropic | $3 | 23.30 | $0.1288 |
| GPT-5.6 Sol(low) | OpenAI | $4.4 | 33.47 | $0.1315 |
| Claude Opus 4.8(max) | Anthropic | $5.5 | 41.79 | $0.1316 |
| Claude Opus 4.7(max) | Anthropic | $5.5 | 40.69 | $0.1352 |
| Claude Opus 5(low) | Anthropic | $5.5 | 39.35 | $0.1398 |
| GPT 5.5(xhigh) | OpenAI | $5.5 | 38.36 | $0.1434 |
| Claude Sonnet 4.5(reasoning) | Anthropic | $3 | 20.67 | $0.1451 |
| GPT 5.5(high) | OpenAI | $5.5 | 36.98 | $0.1487 |
| GPT 5.4(non-reasoning) | OpenAI | $2.75 | 18.16 | $0.1514 |
| Claude Sonnet 4.5(non-reasoning) | Anthropic | $3 | 19.34 | $0.1551 |
| GPT-5.6 Sol(non-reasoning) | OpenAI | $4.4 | 28.33 | $0.1553 |
| Claude Opus 4.6(max) | Anthropic | $5 | 31.95 | $0.1565 |
| GPT 5.5(medium) | OpenAI | $5.5 | 33.80 | $0.1627 |
| Mistral Small (24.02)(Sep '24) | Mistral AI | $1 | 5.85 | $0.1709 |
| Claude Opus 4.5(reasoning) | Anthropic | $5 | 29.10 | $0.1718 |
| Claude Opus 4.7(high) | Anthropic | $5.5 | 30.93 | $0.1778 |
| GPT 5.5(low) | OpenAI | $5.5 | 30.71 | $0.1791 |
| Mistral Small (24.02)(Feb '24) | Mistral AI | $1 | 5.50 | $0.1818 |
| Claude Opus 4.6(high) | Anthropic | $5 | 26.35 | $0.1898 |
| Claude Fable 5.1(max) | Anthropic | $11 | 53.35 | $0.2062 |
| Claude Fable 5.1(xhigh) | Anthropic | $11 | 53.20 | $0.2068 |
| GPT-6 Astra(max) | OpenAI | $11 | 52.67 | $0.2088 |
| GPT-6 Astra(xhigh) | OpenAI | $11 | 52.39 | $0.2100 |
| Claude Opus 4.5(non-reasoning) | Anthropic | $5 | 23.68 | $0.2111 |
| Claude Fable 5.1(high) | Anthropic | $11 | 51.15 | $0.2151 |
| GPT-6 Astra(high) | OpenAI | $11 | 50.92 | $0.2160 |
| Claude Fable 5(max) | Anthropic | $11 | 49.63 | $0.2216 |
| GPT-6 Astra(medium) | OpenAI | $11 | 49.57 | $0.2219 |
| Claude Fable 5.1(medium) | Anthropic | $11 | 48.92 | $0.2249 |
| Claude Fable 5.1(low) | Anthropic | $11 | 46.82 | $0.2349 |
| GPT 5.5(non-reasoning) | OpenAI | $5.5 | 23.17 | $0.2374 |
| GPT-6 Astra(low) | OpenAI | $11 | 45.78 | $0.2403 |
| Pixtral Large (25.02) | Mistral AI | $2 | 7.15 | $0.2797 |
| Llama 3 70B Instruct | Meta | $2.65 | 5.45 | $0.4862 |
| Mistral Large (24.02)(Feb '24) | Mistral AI | $4 | 5.76 | $0.6944 |
De los benchmarks a producción: ¿por qué modelos empezar?
Ningún benchmark sintético reemplaza las pruebas con tus propios datos y flujos de trabajo. En producción, la métrica definitiva no es el precio por millón de tokens, sino el costo por tarea exitosa:
- Lo “barato” puede salir caro: Un modelo de bajo costo que falla con frecuencia, requiere reintentos constantes o prompts excesivamente largos para no alucinar, termina costando más tiempo y dinero.
- Lo más potente suele ser innecesario: Usar el modelo más costoso para tareas repetitivas, clasificación o extracción estructurada es pagar de más sin beneficio real.
Cómo usar esta guía: Probar decenas de modelos contra tus propios datos no es viable. Utiliza la frontera de Pareto como tu shortlist de partida: identifica los 2 o 3 modelos que dominan la relación costo-inteligencia para el nivel de complejidad que necesitas y empieza tus pruebas ahí. Si el más eficiente cumple con el estándar de calidad de tu caso de uso, ya encontraste tu ganador; si no, escala al siguiente escalón de la frontera.
Explora el catálogo completo evaluado
La siguiente tabla reúne los 68 modelos evaluados con sus precios oficiales de entrada y salida ($/1M tokens), ventana de contexto y puntuación de inteligencia. Puedes ordenar por cualquier columna o buscar por modelo y proveedor:
| Modelo↑ | Proveedor | Esfuerzo | Contexto | $/1M in | $/1M out | Índice |
|---|---|---|---|---|---|---|
| Claude Fable 5 | Anthropic | max | 1M | $11.00 | $55.00 | 49.6 |
| Claude Fable 5.1 | Anthropic | high | 1M | $11.00 | $55.00 | 51.1 |
| Claude Fable 5.1 | Anthropic | low | 1M | $11.00 | $55.00 | 46.8 |
| Claude Fable 5.1 | Anthropic | max | 1M | $11.00 | $55.00 | 53.4 |
| Claude Fable 5.1 | Anthropic | medium | 1M | $11.00 | $55.00 | 48.9 |
| Claude Fable 5.1 | Anthropic | xhigh | 1M | $11.00 | $55.00 | 53.2 |
| Claude Haiku 4.5 | Anthropic | non-reasoning | 200K | $1.00 | $5.00 | 15.4 |
| Claude Haiku 4.5 | Anthropic | reasoning | 200K | $1.00 | $5.00 | 16.9 |
| Claude Opus 4.5 | Anthropic | non-reasoning | 200K | $5.00 | $25.00 | 23.7 |
| Claude Opus 4.5 | Anthropic | reasoning | 200K | $5.00 | $25.00 | 29.1 |
| Claude Opus 4.6 | Anthropic | max | 1M | $5.00 | $25.00 | 31.9 |
| Claude Opus 4.6 | Anthropic | non-reasoning | 1M | $5.00 | $25.00 | 26.4 |
| Claude Opus 4.7 | Anthropic | max | 1M | $5.50 | $27.50 | 40.7 |
| Claude Opus 4.7 | Anthropic | non-reasoning | 1M | $5.50 | $27.50 | 30.9 |
| Claude Opus 4.8 | Anthropic | max | 1M | $5.50 | $27.50 | 41.8 |
| Claude Opus 5 | Anthropic | high | 1M | $5.50 | $27.50 | 48.1 |
| Claude Opus 5 | Anthropic | low | 1M | $5.50 | $27.50 | 39.4 |
| Claude Opus 5 | Anthropic | max | 1M | $5.50 | $27.50 | 50.8 |
| Claude Opus 5 | Anthropic | medium | 1M | $5.50 | $27.50 | 44.8 |
| Claude Opus 5 | Anthropic | xhigh | 1M | $5.50 | $27.50 | 49.7 |
| Claude Opus 5.5 | Anthropic | high | 1M | $4.00 | $20.00 | 53.6 |
| Claude Opus 5.5 | Anthropic | low | 1M | $4.00 | $20.00 | 42.3 |
| Claude Opus 5.5 | Anthropic | max | 1M | $4.00 | $20.00 | 57.6 |
| Claude Opus 5.5 | Anthropic | medium | 1M | $4.00 | $20.00 | 51.2 |
| Claude Opus 5.5 | Anthropic | xhigh | 1M | $4.00 | $20.00 | 56.0 |
| Claude Sonnet 4.5 | Anthropic | non-reasoning | 200K | $3.00 | $15.00 | 19.3 |
| Claude Sonnet 4.5 | Anthropic | reasoning | 200K | $3.00 | $15.00 | 20.7 |
| Claude Sonnet 4.6 | Anthropic | max | 1M | $3.00 | $15.00 | 30.1 |
| Claude Sonnet 4.6 | Anthropic | non-reasoning | 1M | $3.00 | $15.00 | 23.3 |
| Claude Sonnet 5 | Anthropic | high | 1M | $2.20 | $11.00 | 31.7 |
| Claude Sonnet 5 | Anthropic | low | 1M | $2.20 | $11.00 | 24.3 |
| Claude Sonnet 5 | Anthropic | max | 1M | $2.20 | $11.00 | 38.2 |
| Claude Sonnet 5 | Anthropic | medium | 1M | $2.20 | $11.00 | 28.1 |
| Claude Sonnet 5 | Anthropic | xhigh | 1M | $2.20 | $11.00 | 34.4 |
| Claude Sonnet 5 | Anthropic | non-reasoning | 1M | $2.20 | $11.00 | 23.2 |
| DeepSeek V3.2 | DeepSeek | non-reasoning | 164K | $0.62 | $1.85 | 16.0 |
| DeepSeek V3.2 | DeepSeek | reasoning | 164K | $0.62 | $1.85 | 21.5 |
| DeepSeek-R1 | DeepSeek | — | 128K | $1.35 | $5.40 | 11.4 |
| Devstral 2 123B | Mistral AI | — | 256K | $0.40 | $2.00 | 8.6 |
| Gemma 3 12B IT | Google | — | 128K | $0.09 | $0.29 | 3.8 |
| Gemma 3 27B PT | Google | — | 128K | $0.23 | $0.38 | 4.8 |
| Gemma 3 4B IT | Google | — | 128K | $0.04 | $0.08 | 4.8 |
| GLM 4.7 | Z.AI | non-reasoning | 203K | $0.60 | $2.20 | 17.4 |
| GLM 4.7 | Z.AI | reasoning | 203K | $0.60 | $2.20 | 22.2 |
| GLM 4.7 Flash | Z.AI | non-reasoning | 203K | $0.07 | $0.40 | 10.6 |
| GLM 4.7 Flash | Z.AI | reasoning | 203K | $0.07 | $0.40 | 14.9 |
| GLM 5 | Z.AI | non-reasoning | 200K | $1.00 | $3.20 | 21.8 |
| GLM 5 | Z.AI | reasoning | 200K | $1.00 | $3.20 | 27.9 |
| GPT 5.4 | OpenAI | non-reasoning | 272K(≤272K) | $2.75 | $13.75 | 18.2 |
| GPT 5.4 | OpenAI | non-reasoning | 1M(>272K) | $5.50 | $20.63 | 18.2 |
| GPT 5.4 | OpenAI | low | 272K(≤272K) | $2.75 | $13.75 | 27.6 |
| GPT 5.4 | OpenAI | low | 1M(>272K) | $5.50 | $20.63 | 27.6 |
| GPT 5.4 | OpenAI | xhigh | 272K(≤272K) | $2.75 | $13.75 | 39.0 |
| GPT 5.4 | OpenAI | xhigh | 1M(>272K) | $5.50 | $20.63 | 39.0 |
| GPT 5.5 | OpenAI | non-reasoning | 272K(≤272K) | $5.50 | $33.00 | 23.2 |
| GPT 5.5 | OpenAI | non-reasoning | 1M(>272K) | $11.00 | $49.50 | 23.2 |
| GPT 5.5 | OpenAI | low | 272K(≤272K) | $5.50 | $33.00 | 30.7 |
| GPT 5.5 | OpenAI | low | 1M(>272K) | $11.00 | $49.50 | 30.7 |
| GPT 5.5 | OpenAI | medium | 272K(≤272K) | $5.50 | $33.00 | 33.8 |
| GPT 5.5 | OpenAI | medium | 1M(>272K) | $11.00 | $49.50 | 33.8 |
| GPT 5.5 | OpenAI | high | 272K(≤272K) | $5.50 | $33.00 | 37.0 |
| GPT 5.5 | OpenAI | high | 1M(>272K) | $11.00 | $49.50 | 37.0 |
| GPT 5.5 | OpenAI | xhigh | 272K(≤272K) | $5.50 | $33.00 | 38.4 |
| GPT 5.5 | OpenAI | xhigh | 1M(>272K) | $11.00 | $49.50 | 38.4 |
| GPT-5.6 Luna | OpenAI | non-reasoning | 272K(≤272K) | $0.22 | $1.32 | 15.5 |
| GPT-5.6 Luna | OpenAI | non-reasoning | 1M(>272K) | $0.44 | $1.98 | 15.5 |
| GPT-5.6 Luna | OpenAI | high | 272K(≤272K) | $0.22 | $1.32 | 32.1 |
| GPT-5.6 Luna | OpenAI | high | 1M(>272K) | $0.44 | $1.98 | 32.1 |
| GPT-5.6 Luna | OpenAI | low | 272K(≤272K) | $0.22 | $1.32 | 21.0 |
| GPT-5.6 Luna | OpenAI | low | 1M(>272K) | $0.44 | $1.98 | 21.0 |
| GPT-5.6 Luna | OpenAI | max | 272K(≤272K) | $0.22 | $1.32 | 37.3 |
| GPT-5.6 Luna | OpenAI | max | 1M(>272K) | $0.44 | $1.98 | 37.3 |
| GPT-5.6 Luna | OpenAI | medium | 272K(≤272K) | $0.22 | $1.32 | 25.0 |
| GPT-5.6 Luna | OpenAI | medium | 1M(>272K) | $0.44 | $1.98 | 25.0 |
| GPT-5.6 Luna | OpenAI | xhigh | 272K(≤272K) | $0.22 | $1.32 | 34.6 |
| GPT-5.6 Luna | OpenAI | xhigh | 1M(>272K) | $0.44 | $1.98 | 34.6 |
| GPT-5.6 Sol | OpenAI | non-reasoning | 272K(≤272K) | $4.40 | $22.00 | 28.3 |
| GPT-5.6 Sol | OpenAI | non-reasoning | 1M(>272K) | $8.80 | $33.00 | 28.3 |
| GPT-5.6 Sol | OpenAI | high | 272K(≤272K) | $4.40 | $22.00 | 42.4 |
| GPT-5.6 Sol | OpenAI | high | 1M(>272K) | $8.80 | $33.00 | 42.4 |
| GPT-5.6 Sol | OpenAI | low | 272K(≤272K) | $4.40 | $22.00 | 33.5 |
| GPT-5.6 Sol | OpenAI | low | 1M(>272K) | $8.80 | $33.00 | 33.5 |
| GPT-5.6 Sol | OpenAI | max | 272K(≤272K) | $4.40 | $22.00 | 47.0 |
| GPT-5.6 Sol | OpenAI | max | 1M(>272K) | $8.80 | $33.00 | 47.0 |
| GPT-5.6 Sol | OpenAI | medium | 272K(≤272K) | $4.40 | $22.00 | 39.2 |
| GPT-5.6 Sol | OpenAI | medium | 1M(>272K) | $8.80 | $33.00 | 39.2 |
| GPT-5.6 Sol | OpenAI | xhigh | 272K(≤272K) | $4.40 | $22.00 | 44.0 |
| GPT-5.6 Sol | OpenAI | xhigh | 1M(>272K) | $8.80 | $33.00 | 44.0 |
| GPT-5.6 Terra | OpenAI | non-reasoning | 272K(≤272K) | $2.20 | $13.20 | 20.8 |
| GPT-5.6 Terra | OpenAI | non-reasoning | 1M(>272K) | $4.40 | $19.80 | 20.8 |
| GPT-5.6 Terra | OpenAI | high | 272K(≤272K) | $2.20 | $13.20 | 34.2 |
| GPT-5.6 Terra | OpenAI | high | 1M(>272K) | $4.40 | $19.80 | 34.2 |
| GPT-5.6 Terra | OpenAI | low | 272K(≤272K) | $2.20 | $13.20 | 27.5 |
| GPT-5.6 Terra | OpenAI | low | 1M(>272K) | $4.40 | $19.80 | 27.5 |
| GPT-5.6 Terra | OpenAI | max | 272K(≤272K) | $2.20 | $13.20 | 42.1 |
| GPT-5.6 Terra | OpenAI | max | 1M(>272K) | $4.40 | $19.80 | 42.1 |
| GPT-5.6 Terra | OpenAI | medium | 272K(≤272K) | $2.20 | $13.20 | 30.1 |
| GPT-5.6 Terra | OpenAI | medium | 1M(>272K) | $4.40 | $19.80 | 30.1 |
| GPT-5.6 Terra | OpenAI | xhigh | 272K(≤272K) | $2.20 | $13.20 | 38.0 |
| GPT-5.6 Terra | OpenAI | xhigh | 1M(>272K) | $4.40 | $19.80 | 38.0 |
| GPT-6 Astra | OpenAI | high | 272K(≤272K) | $11.00 | $55.00 | 50.9 |
| GPT-6 Astra | OpenAI | high | 1.1M(>272K) | $22.00 | $82.50 | 50.9 |
| GPT-6 Astra | OpenAI | low | 272K(≤272K) | $11.00 | $55.00 | 45.8 |
| GPT-6 Astra | OpenAI | low | 1.1M(>272K) | $22.00 | $82.50 | 45.8 |
| GPT-6 Astra | OpenAI | max | 272K(≤272K) | $11.00 | $55.00 | 52.7 |
| GPT-6 Astra | OpenAI | max | 1.1M(>272K) | $22.00 | $82.50 | 52.7 |
| GPT-6 Astra | OpenAI | medium | 272K(≤272K) | $11.00 | $55.00 | 49.6 |
| GPT-6 Astra | OpenAI | medium | 1.1M(>272K) | $22.00 | $82.50 | 49.6 |
| GPT-6 Astra | OpenAI | xhigh | 272K(≤272K) | $11.00 | $55.00 | 52.4 |
| GPT-6 Astra | OpenAI | xhigh | 1.1M(>272K) | $22.00 | $82.50 | 52.4 |
| GPT-6 Luna | OpenAI | non-reasoning | 272K(≤272K) | $0.11 | $0.55 | 18.3 |
| GPT-6 Luna | OpenAI | non-reasoning | 1M(>272K) | $0.22 | $0.82 | 18.3 |
| GPT-6 Luna | OpenAI | high | 272K(≤272K) | $0.11 | $0.55 | 32.1 |
| GPT-6 Luna | OpenAI | high | 1M(>272K) | $0.22 | $0.82 | 32.1 |
| GPT-6 Luna | OpenAI | low | 272K(≤272K) | $0.11 | $0.55 | 20.9 |
| GPT-6 Luna | OpenAI | low | 1M(>272K) | $0.22 | $0.82 | 20.9 |
| GPT-6 Luna | OpenAI | max | 272K(≤272K) | $0.11 | $0.55 | 37.3 |
| GPT-6 Luna | OpenAI | max | 1M(>272K) | $0.22 | $0.82 | 37.3 |
| GPT-6 Luna | OpenAI | medium | 272K(≤272K) | $0.11 | $0.55 | 29.5 |
| GPT-6 Luna | OpenAI | medium | 1M(>272K) | $0.22 | $0.82 | 29.5 |
| GPT-6 Luna | OpenAI | xhigh | 272K(≤272K) | $0.11 | $0.55 | 33.9 |
| GPT-6 Luna | OpenAI | xhigh | 1M(>272K) | $0.22 | $0.82 | 33.9 |
| GPT-6 Sol | OpenAI | non-reasoning | 272K(≤272K) | $2.20 | $11.00 | 28.1 |
| GPT-6 Sol | OpenAI | non-reasoning | 1M(>272K) | $4.40 | $16.50 | 28.1 |
| GPT-6 Sol | OpenAI | high | 272K(≤272K) | $2.20 | $11.00 | 42.8 |
| GPT-6 Sol | OpenAI | high | 1M(>272K) | $4.40 | $16.50 | 42.8 |
| GPT-6 Sol | OpenAI | low | 272K(≤272K) | $2.20 | $11.00 | 33.9 |
| GPT-6 Sol | OpenAI | low | 1M(>272K) | $4.40 | $16.50 | 33.9 |
| GPT-6 Sol | OpenAI | max | 272K(≤272K) | $2.20 | $11.00 | 47.5 |
| GPT-6 Sol | OpenAI | max | 1M(>272K) | $4.40 | $16.50 | 47.5 |
| GPT-6 Sol | OpenAI | medium | 272K(≤272K) | $2.20 | $11.00 | 39.8 |
| GPT-6 Sol | OpenAI | medium | 1M(>272K) | $4.40 | $16.50 | 39.8 |
| GPT-6 Sol | OpenAI | xhigh | 272K(≤272K) | $2.20 | $11.00 | 44.1 |
| GPT-6 Sol | OpenAI | xhigh | 1M(>272K) | $4.40 | $16.50 | 44.1 |
| gpt-oss-120b | OpenAI | high | 128K | $0.15 | $0.60 | 11.6 |
| gpt-oss-120b | OpenAI | low | 128K | $0.15 | $0.60 | 10.2 |
| gpt-oss-20b | OpenAI | high | 128K | $0.07 | $0.30 | 9.0 |
| gpt-oss-20b | OpenAI | low | 128K | $0.07 | $0.30 | 9.9 |
| Grok 4.6 | xAI | high | 500K | $2.20 | $6.60 | 44.3 |
| Grok 4.6 | xAI | low | 500K | $2.20 | $6.60 | 35.1 |
| Grok 4.6 | xAI | medium | 500K | $2.20 | $6.60 | 42.8 |
| Grok 4.6 | xAI | xhigh | 500K | $2.20 | $6.60 | 44.2 |
| Kimi K2 Thinking | Moonshot AI | — | 256K | $0.60 | $2.50 | 22.0 |
| Kimi K2.5 | Moonshot AI | non-reasoning | 256K | $0.60 | $3.00 | 19.4 |
| Kimi K2.5 | Moonshot AI | reasoning | 256K | $0.60 | $3.00 | 23.5 |
| Kimi K3 | Moonshot AI | low | 1M | $3.30 | $16.50 | 34.5 |
| Kimi K3 | Moonshot AI | max | 1M | $3.30 | $16.50 | 43.6 |
| Llama 3 70B Instruct | Meta | — | 8K | $2.65 | $3.50 | 5.5 |
| Llama 3 8B Instruct | Meta | — | 8K | $0.30 | $0.60 | 4.8 |
| Llama 3.1 70B Instruct | Meta | — | 128K | $0.72 | $0.72 | 6.6 |
| Llama 3.1 8B Instruct | Meta | — | 128K | $0.22 | $0.22 | 6.9 |
| Llama 3.3 70B Instruct | Meta | — | 128K | $0.72 | $0.72 | 7.7 |
| Llama 4 Maverick 17B Instruct | Meta | — | 1M | $0.24 | $0.97 | 10.0 |
| Llama 4 Scout 17B Instruct | Meta | — | 10M | $0.17 | $0.66 | 8.1 |
| Magistral Small 2509 | Mistral AI | — | 128K | $0.50 | $1.50 | 8.6 |
| MiniMax M2 | MiniMax | — | 1M | $0.30 | $1.20 | 18.6 |
| MiniMax M2.1 | MiniMax | — | 196K | $0.30 | $1.20 | 20.9 |
| MiniMax M2.5 | MiniMax | — | 196K | $0.30 | $1.20 | 22.8 |
| Ministral 14B 3.0 | Mistral AI | — | 128K | $0.20 | $0.20 | 6.0 |
| Ministral 3 8B | Mistral AI | — | 128K | $0.15 | $0.15 | 5.5 |
| Ministral 3B | Mistral AI | — | 128K | $0.10 | $0.10 | 4.8 |
| Mistral 7B Instruct | Mistral AI | — | 32K | $0.15 | $0.20 | 5.0 |
| Mistral Large (24.02) | Mistral AI | — | 32K | $4.00 | $12.00 | 5.8 |
| Mistral Large 3 | Mistral AI | — | 256K | $0.50 | $1.50 | 9.3 |
| Mistral Small (24.02) | Mistral AI | — | 32K | $1.00 | $3.00 | 5.8 |
| Mixtral 8x7B Instruct | Mistral AI | — | 32K | $0.45 | $0.70 | 5.1 |
| Nemotron Nano 3 30B | NVIDIA | — | 256K | $0.06 | $0.24 | 6.8 |
| Nova 2 Lite | Amazon | non-reasoning | 1M | $0.30 | $2.50 | 8.7 |
| Nova 2 Lite | Amazon | high | 1M | $0.30 | $2.50 | 13.4 |
| Nova 2 Lite | Amazon | low | 1M | $0.30 | $2.50 | 11.8 |
| Nova 2 Lite | Amazon | medium | 1M | $0.30 | $2.50 | 12.5 |
| Nova Lite | Amazon | — | 300K | $0.06 | $0.24 | 6.7 |
| Nova Micro | Amazon | — | 128K | $0.035 | $0.14 | 5.9 |
| Nova Pro | Amazon | — | 300K | $0.80 | $3.20 | 7.0 |
| NVIDIA Nemotron 3 Super 120B A12B | NVIDIA | — | 256K | $0.15 | $0.65 | 12.8 |
| NVIDIA Nemotron Nano 12B v2 VL BF16 | NVIDIA | — | 128K | $0.20 | $0.60 | 5.8 |
| NVIDIA Nemotron Nano 9B v2 | NVIDIA | non-reasoning | 128K | $0.06 | $0.23 | 6.8 |
| NVIDIA Nemotron Nano 9B v2 | NVIDIA | reasoning | 128K | $0.06 | $0.23 | 7.4 |
| Pixtral Large (25.02) | Mistral AI | — | 128K | $2.00 | $6.00 | 7.2 |
| Qwen3 32B (dense) | Qwen | non-reasoning | 32K | $0.15 | $0.60 | 7.3 |
| Qwen3 32B (dense) | Qwen | reasoning | 32K | $0.15 | $0.60 | 8.6 |
| Qwen3 Coder Next | Qwen | — | 256K | $0.50 | $1.20 | 9.2 |
| Qwen3 Next 80B A3B | Qwen | reasoning | 256K | $0.15 | $1.20 | 11.2 |
| Qwen3 VL 235B A22B | Qwen | reasoning | 256K | $0.53 | $2.66 | 13.4 |
| Qwen3-Coder-30B-A3B-Instruct | Qwen | — | 256K | $0.15 | $0.60 | 9.6 |
185 de 185 configuraciones (68 modelos únicos) · clic en un encabezado para ordenar


