---
name: Qwen3.5 397B-A17B
slug: qwen-3-5-397b-a17b
type: Multimodal
providers:
  - qiniu-ai
  - togetherai
  - chutes
  - nebius
  - siliconflow-cn
  - ovhcloud
  - nano-gpt
  - huggingface
  - nvidia
  - neuralwatt
  - scaleway
  - novita-ai
  - merge-gateway
  - alibaba-cn
  - infomaniak
  - mixlayer
  - alibaba
  - openrouter
description: 'Qwen vision-language model for visual reasoning, documents, and agent tasks'
tags:
  - open-source
  - reasoning
  - tool-use
  - multimodal
  - qwen
is_open_source: true
context_window: 262144
max_output: 262144
pricing:
  - image: 0
    input: 0.43
    output: 2.58
    provider: alibaba-cn
  - image: 0
    input: 0.6
    output: 3.6
    provider: alibaba
  - image: 0
    input: 0.45
    output: 3
    provider: chutes (TEE)
  - image: 0
    input: 0.6
    output: 3.6
    provider: huggingface
  - image: 0
    input: 0.99
    output: 4.46
    provider: infomaniak (FP8)
  - image: 0
    input: 0.172
    output: 1.032
    provider: merge-gateway
  - image: 0
    input: 0.6
    output: 3.6
    provider: mixlayer
  - image: 0
    input: 0.55
    output: 3.5
    provider: nano-gpt (TEE)
  - image: 0
    input: 0.6
    output: 3.6
    provider: nano-gpt
  - image: 0
    input: 0.6
    output: 3.6
    provider: nebius
  - image: 0
    input: 0.69
    output: 4.14
    provider: neuralwatt (FP8)
  - image: 0
    input: 0.6
    output: 3.6
    provider: novita-ai
  - image: 0
    input: 0
    output: 0
    provider: nvidia
  - image: 0
    input: 0.71
    output: 4.25
    provider: ovhcloud
  - image: 0
    input: 0.6
    output: 3.6
    provider: scaleway
  - image: 0
    input: 0.29
    output: 1.74
    provider: siliconflow-cn
  - image: 0
    input: 0.6
    output: 3.6
    provider: togetherai
  - image: 0
    input: 0.5
    output: 3.5999999999999996
    provider: openrouter
release_date: '2026-02-16'
openrouter_id: qwen/qwen3.5-397b-a17b
capabilities:
  audio: true
  vision: true
  tool_use: true
  citations: false
  reasoning: true
---
Qwen3.5 397B-A17B is a Multimodal model.
Qwen vision-language model for visual reasoning, documents, and agent tasks