Qwen3.8 Omni Flash

Qwen3.8 Omni Flash is an omni model from Alibaba's Qwen line for text, vision, audio, and multimodal agent tasks. It is the successor to Qwen3.8 Flash, released 2026-08-26. It has a context window of 1,000,000 tokens and a max output of 131,072 tokens. Pricing starts at $0.1126 per million input tokens and $0.38 per million output tokens, with cached input at $0.014 per million. Weights are not open. It accepts audio, image, text, and video, and returns text. It supports reasoning mode and tool calling.

Against Qwen3.8 Flash: input $0.11 to $0.1126, output $0.38 to $0.38, cached input $0.011 to $0.014.

Range across the hosts that serve it: up to $0.15 per 1M input tokens.

• ·Handles text, vision, audio, and video inputs
• ·1M token context window
• ·131K max output tokens
• ·Supports reasoning and tool calling
• ·Served by 9 hosts

• Reach for it for multimodal agent tasks
• Reach for it for long-context processing up to 1M tokens
• Reach for it for audio and video understanding
• Reach for it for reasoning with tool calling

9 hosts serve this model at the price above · the maker's documentation

• 8Qwen 发布原生全模态模型 Qwen3.8-Omni-Flash，主打音视频智能体任务交付2026-09-18

1 signal of the archive mention this model, each scored against the published bar. This list is rebuilt from the archive at every build, so it cannot go stale.

The registry keeps the six most recent versions of a line. Older ones stay in the file and can be surfaced without a rebuild.

What is the context window?

The context window is 1,000,000 tokens.

Are the weights open?

No, the weights are not open.

What input types does it accept?

It accepts audio, image, text, and video.