← All models

Qwen3-VL-Flash

Published Publisher Alibabaqwen3-vl-flash

Overview

Qwen3-VL-Flash is a small, fast-response vision-language model in the Qwen3 series. It combines thinking and non-thinking modes. The specified snapshot is qwen3-vl-flash-2025-10-15. (help.aliyun.com)

Capabilities and uses

  • Accepts text, images and video, and produces text. (qwencloud.com)
  • Supports visual understanding, reasoning, object recognition, spatial awareness and 2D/3D visual localization. Alibaba describes support for long documents and videos. Suitable uses include document and video analysis and visual question answering. (help.aliyun.com)

Release and availability

Alibaba's release record lists the model and snapshot under October 14, 2025 for its domestic listing and October 15, 2025 for international deployment. The snapshot name also carries the October 15 date. (help.aliyun.com)

Alibaba's domestic retirement notice listed this snapshot for retirement on October 10, 2026, at 00:00 Beijing time, subject to the actual change time. Verify availability in the intended region before integrating it. The undated API ID is an alias and can change: Alibaba announced that its Beijing and Singapore mapping would change from the 2025 snapshot to qwen3-vl-flash-2026-01-22 beginning March 24, 2026. (aliyun.com)

Limitations

The model outputs text rather than generated images or video. Visual recognition and reasoning can be inaccurate, so important results should be checked. Availability and exact input and output limits depend on the selected version and service region. (qwencloud.com)

Pricing

Checked November 17, 2026. The figures below are historical official rates for the specified 2025 snapshot, not a guarantee of current availability or pricing. The current pricing pages no longer list the snapshot; check the relevant console before use. Rates are per 1 million tokens, separately for input and output; the tier is determined by input tokens per request. Output includes reasoning and answer tokens.

Official service and scope Input tokens per request Input Output
Mainland Alibaba Cloud, Beijing 0–32K CNY 0.15 CNY 1.5
Mainland Alibaba Cloud, Beijing >32K–128K CNY 0.3 CNY 3
Mainland Alibaba Cloud, Beijing >128K–256K CNY 0.6 CNY 6
Alibaba Cloud international site, Virginia, Global deployment 0–32K USD 0.022 USD 0.215
Alibaba Cloud international site, Virginia, Global deployment >32K–128K USD 0.043 USD 0.43
Alibaba Cloud international site, Virginia, Global deployment >128K–256K USD 0.086 USD 0.859
Alibaba Cloud international site, Virginia, US deployment 0–32K USD 0.05 USD 0.4
Alibaba Cloud international site, Virginia, US deployment >32K–128K USD 0.075 USD 0.6
Alibaba Cloud international site, Virginia, US deployment >128K–256K USD 0.12 USD 0.96

These are standard rates for thinking and non-thinking modes, not cache-hit or batch discounts. Regional deployments and domestic and international accounts are distinct offerings; the currencies have not been converted. Historical official pricing sources: mainland pricing and international pricing. (help.aliyun.com)

Official sources