Qwen3.8-Max — Specifications
Compilado por Mustafa Ihsan de la documentación publicada por el proveedor · Última actualización
| Desarrollador | Alibaba |
|---|---|
| Tipo | LLM (mixture of experts, razonamiento, multimodal) |
| Modalidad | Text, Image, Video → Text |
| Parámetros | 2.4T total (MoE) |
| Ventana de contexto | 1 millón |
| Salida máxima | 128 K |
| Licencia | Proprietary (hosted) |
| Pesos abiertos | No |
| Lanzado | 2026-08 |
| Precio de entrada | $2.00 /1M |
| Precio de salida | $6.00 /1M |
| Proveedores de API | Alibaba Cloud Model Studio, Qwen Cloud |
What is Qwen3.8-Max?
Qwen3.8-Max is Alibaba’s flagship Qwen model, launched on Alibaba Cloud Model Studio (International) on 2 August 2026. Alibaba lo describe as a 2.4-trillion-parameter mixture-of-experts with a major leap in coding and office productivity, built to handle professional tasks in law, finance and design, with native understanding of images and video. Qwen calls it the official version of its open-weight Qwen3.8-2.4T-A95B (2.4T total, 95B active parameters), with extra features such as vision input, a non-thinking mode, a 1M-token context by default and built-in tools. An upgraded snapshot, qwen3.8-max-0902, followed on 2 September 2026 at the same price.
In Model Studio’s US and International deployment scopes it costs $2.00 per million input tokens and $6.00 per million output tokens, with implicit cache hits at $0.25 per million. Model Studio’s Global deployment scope lists it at $1.65 / $4.951, and Qwen Cloud, the official Qwen API service, charges $2.00 / $6.00. The context window is 1,000,000 tokens, with up to 131,072 output tokens.
It succeeds the April 2025 Qwen3 generation at the top of Alibaba’s range; the open Qwen3 235B-A22B remains the self-hostable flagship of that family, with a 32K native context. For Alibaba’s smaller open-weight Qwen3.8 model, see Qwen3.8-27B. Background on the family is in our Explicación de Qwen de Alibaba guide, and the Calculadora de costos de API de IA estimates a monthly bill.
Qwen3.8-Max pricing: API cost per 1M tokens
| Entrada (por cada millón de tokens) | $2.00 |
|---|---|
| Salida (por cada millón de tokens) | $6.00 |
| Relación salida/entrada | 3× |
| Combinada (4:1 entrada:salida) | $2.80 por 1 millón de tokens |
What Qwen3.8-Max costs per month
Gasto mensual real con una mezcla entrada:salida de 4:1, es decir, la proporción que realmente genera una carga de trabajo típica de chat o RAG.
| Carga de trabajo | Tokens/mes | Coste mensual |
|---|---|---|
| Proyecto secundario | 1 millón de tokens de entrada / 0,25 millones de tokens de salida | $3.50 |
| Pequeño equipo | 20 millones de tokens de entrada / 5 millones de tokens de salida | $70 |
| Producción | 200 millones de tokens de entrada / 50 millones de tokens de salida | $700 |
Calcule sus propios números en la Calculadora de costos de API de IA.
Cheaper alternatives to Qwen3.8-Max
| Modelo | Dólares por millón combinados | Usted ahorra |
|---|---|---|
| Llama 3.1 8B abierta | $0.0240 | 99 % más barato |
| Mistral NeMo 12B abierta | $0.0292 | 99 % más barato |
| Gemma 3 4B abierta | $0.0600 | 98% cheaper |
Preguntas frecuentes
How much does Qwen3.8-Max cost per 1M tokens?
Qwen3.8-Max costs $2.00 per 1M input tokens and $6.00 per 1M output tokens. At a typical 4:1 input-to-output mix that blends to about $2.80 per 1M tokens.
How much does Qwen3.8-Max cost per month?
A small-team workload of 20M input and 5M output tokens a month costs about $70 on Qwen3.8-Max. A side project (1M in / 0.25M out) costs roughly $3.50.
What is a cheaper alternative to Qwen3.8-Max?
Llama 3.1 8B is the strongest cheaper option in our database at $0.0240 per 1M blended — about 99% less than Qwen3.8-Max. It is also open-weight, so self-hosting is an option.
Can I run Qwen3.8-Max locally?
No. Qwen3.8-Max is a closed, API-only model — the weights are not released, so it cannot be self-hosted.
Why does Qwen3.8-Max charge more for output than input?
Output tokens are generated one at a time and cannot be batched the way a prompt can, so they cost the provider more to serve. Qwen3.8-Max charges 3× more for output, which is why prompt-heavy workloads are far cheaper to run than generation-heavy ones.
Los precios corresponden a las tarifas oficiales publicadas para la API principal del modelo y se revisan periódicamente conforme los proveedores los actualicen. No incluyen descuentos por volumen, procesamiento por lotes ni entradas en caché. Compare todos los modelos uno al lado del otro en la Base de datos de modelos de IA o el Clasificación de modelos de lenguaje grande (LLM).