Model detail
qwen3.5-397b-a17b
The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers state-of-the-art performance comparable to leading-edge models across a wide range of tasks, including language understanding, logical reasoning, code generation, agent-based tasks, image understanding, video understanding, and graphical user interface (GUI) interactions. With its robust code-generation and agent capabilities, the model exhibits strong generalization across diverse agent.
Model specs
- Model ID
-
qwen3.5-397b-a17b
- Context length
- 262.144K
- Max output
- 65.536K
- I/O modalities
- Released
- 2026-02
02
API endpoints
One MixRoute gateway, OpenAI-compatible
-
OpenAI-compatible
/v1/chat/completionsPOST
03
Pricing
Flat rate, unit: /1M Tokens
Input
$0.6000 /1M Tokens
Completion
$3.6000 /1M Tokens
05
Selection summary
Quickly judge whether it fits your workload
What is this model good for?
The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid
The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency.
What should you check before using it?
262,144 context, 65,536 output tokens
Alibaba model that supports reasoning, tool calling and structured output. Released 2026-02-15.
Why use it through MixRoute?
qwen3.5-397b-a17b takes text and images on one route
Text and image inputs share the same route and the same key. MixRoute serves qwen3.5-397b-a17b on the OpenAI-compatible endpoint /v1/chat/completions; point your existing client at https://api.mixroute.ai/v1 with the model ID qwen3.5-397b-a17b and re-run your integration tests.
07
Capabilities & compatibility
Unconfirmed capabilities stay "Unknown," no guesses
Tool calling
qwen3.5-397b-a17b supports function and tool calls.
tools definition; the response returns tool_calls to execute and feed results back.
Structured output
qwen3.5-397b-a17b can produce JSON-formatted output on demand.
response_format: { type: "json_schema" } forces valid JSON output.
Streaming
qwen3.5-397b-a17b supports streamed responses.
stream: true streams tokens incrementally, so chat UIs render as they arrive.
Vision input
qwen3.5-397b-a17b can process image inputs alongside text.
image_url content part alongside text messages.
OpenAI SDK
Call qwen3.5-397b-a17b through MixRoute's OpenAI-compatible endpoint.
base_url to https://api.mixroute.ai/v1; everything else follows the OpenAI style.
08
Token cost estimator
Live estimate from this page's pricing, not an actual bill
This model has no cache-read price; caching is not counted
09
FAQ
qwen3.5-397b-a17b is billed per use and the record publishes no free-tier allowance. The current published rates are input $0.6000 per 1M tokens and output $3.6000 per 1M tokens. Trial credit or promotional quota is tied to your account rather than to the model, so check the console before planning around a free tier.
One endpoint, a testable decision
Test this model and alternate routes with the same request format
Start from a real workload, then let quality, total cost, and failure conditions decide whether to send production traffic.