Qwen3.8-Max

Qwen3.8-Max is QwenCloud's stable hosted 2.4-trillion-parameter multimodal model; its September 2 snapshot is qwen3.8-max-0902.

Language ModelMultimodalLong ContextQwenCloud

QwenCloud or Alibaba Cloud Model Studio hosted API

$2/1M input, $0.25/1M implicit-cache input, and $6/1M output tokens on QwenCloud

Commercial

🚀Function Overview

A hosted frontier model for long-context multimodal reasoning and agent workflows where a stable API ID, published limits, and token pricing are required.

Key Features

Stable request model ID qwen3.8-max
September 2 snapshot qwen3.8-max-0902 with dated alias qwen3.8-max-2026-09-02
Vendor-reported snapshot gains in coding, long-horizon agents, tool orchestration, chart reasoning, document parsing, and multimodal perception
Image, text, and video input with text output
1,000,000-token context window
Up to 991,000 input tokens in non-thinking mode or 983,000 in thinking mode
131,000-token maximum output and up to 262,000 reasoning tokens
Function calling, structured output, batches, web search, caching, and fine-tuning
$2 input, $0.25 implicit-cache input, and $6 output per 1 million tokens on QwenCloud

Use Cases

•Analyzing long documents, repositories, or mixed-media context
•Running coding and tool-using agent workflows through a stable hosted API
•Reasoning over image or video inputs with a text response
•Comparing frontier-model quality and cost with a fixed evaluation harness

⚙️Input Parameters

messages

array

OpenAI-compatible messages containing text and supported image or video content.

model

string

Use the stable qwen3.8-max request model ID.

reasoning_effort

string

Select a supported reasoning mode and budget for the task.

💡Usage Examples

Example 1

Input Parameters

{
  "model": "qwen3.8-max",
  "messages": [
    {
      "role": "user",
      "content": "Compare these two architecture proposals and identify the smallest safe migration plan."
    }
  ]
}

Output Results

A QwenCloud response from Qwen3.8-Max. Validate quality, latency, and total cost on your own workload before production routing.

Quick Actions

Technical Specifications

Hardware Type: QwenCloud or Alibaba Cloud Model Studio hosted API
Commercial Use: Supported
Pricing: $2/1M input, $0.25/1M implicit-cache input, and $6/1M output tokens on QwenCloud