Qwen 3.8 Max-0902 API

Capabilities

Context length

1000000 tokens

Max output

131072 tokens

Official snapshot alias

qwen3.8-max-2026-09-02

Maximum input — non-thinking

991,808 tokens

Maximum input — thinking

983,616 tokens

Maximum reasoning length

262,144 tokens

Batch inference

Not supported

Fine-tuning

Not supported

• Coding
• Tool orchestration
• Document analysis
• reasoning
• function_calling
• structured_output
• caching
• vision
• prefix_completion
• web_search

API pricing

Input

$1.6 USD per 1M tokens

Output

$4.800000000000001 USD per 1M tokens

Cached input

$0.2 USD per 1M tokens

About this snapshot

Qwen 3.8 Max-0902 is Alibaba Cloud’s September 2 snapshot. Its updates focus on coding, coordinated agents, and visual understanding. Use the exact model ID when you want to select this dated version.

Core strengths

Coding

Complex projects and sustained development tasks.

Agents

Planning and coordinating multiple tools.

Visual understanding

Chart reasoning and document parsing.

Example workflows

Start with a small, representative task and evaluate the response before expanding your workflow.

Code review

Provide a change and its requirements; ask for a review with specific fixes.

Document extraction

Define a target schema, supply source material, and validate the returned fields.

Tool-assisted tasks

Define your tools, execute returned calls in your application, and send the results back.

Specifications and limits

The values in Capabilities are Alibaba Cloud’s published model limits. They are not additive: leave room for the response within each request’s context budget. Account and API configuration can impose lower limits.

Get started with BulkToken

Use a BulkToken API key and the exact model ID shown on this page.

Prepare your key

Create an API key in the console and make sure your account has credits. Store the key in the BULKTOKEN_API_KEY environment variable.

Configure your client

Set the base URL to https://bulktoken.ai/v1 and the model to qwen3.8-max-0902.

Send a small request

Use the example below, inspect the returned message and usage, then review the request in Usage Logs.

API examples

These examples use the OpenAI-compatible Chat Completions endpoint. For Python, install the openai package first. Set BULKTOKEN_API_KEY in your environment before running either example.

Python

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["BULKTOKEN_API_KEY"],
    base_url="https://bulktoken.ai/v1",
)

response = client.chat.completions.create(
    model="qwen3.8-max-0902",
    messages=[{"role": "user", "content": "Explain what an API gateway does in three sentences."}],
    max_tokens=256,
)

print(response.choices[0].message.content)
print(response.usage)

cURL

curl https://bulktoken.ai/v1/chat/completions \
  -H "Authorization: Bearer $BULKTOKEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.8-max-0902",
    "messages": [{"role": "user", "content": "Explain what an API gateway does in three sentences."}],
    "max_tokens": 256
  }'

FAQ

Which model name should I send?

Use qwen3.8-max-0902 exactly as shown. A display name is not an API identifier.

How is the discount applied?

The displayed BulkToken rates already include the 20% discount. They are USD per 1 million tokens; do not apply the discount a second time.

What does the cache-read rate cover?

The listed cache-read rate covers implicit cache hits. Automatic caching does not guarantee a hit. Explicit caching has distinct creation and read rates; confirm its pricing before using that mode.

Official references

Source documents used for this model listing.

Qwen 3.8 Max — snapshot specifications

Official model and snapshot documentation.

Alibaba Cloud model pricing

Official list prices and billing categories.

Context caching

Implicit and explicit caching behavior.