Qwen 3.6 Flash

Qwen[text]

Cheapest video-capable model with massive 1M context window

Input

$0.10

per 1M tokens

Output

$0.40

per 1M tokens

Context

1M tokens

Max Output

66K tokens

Released

2026-05-04

Features

[native-video-input][1m-tokens-context][ultra-cheap][multimodal-reasoning]

Technical Specifications

The Lab

location: Hangzhou, China (Alibaba DAMO)founded: 2023backed by: Alibaba Group (DAMO Academy / Tongyi Lab)

Qwen is the open-weights language-model brand of Alibaba's DAMO Academy / Tongyi Lab, headquartered in Hangzhou. The series gained wide adoption starting with Qwen2 in 2024 as a competitive open-source alternative to Llama, with strong multilingual and coding performance. By 2026 the family includes frontier reasoning models, multimodal omni models, and code-specialised variants.

more about Qwen →

Model ID: qwen-3.6-flash