Pyyan / Language Models / Qwen3.8-Omni-Flash

Qwen3.8-Omni-Flash

Text, image, audio and video into one million tokens of context.

currentAlibaba ↗Verified 2026-09-20

$0.15

per M input

Specifications

Context1Msource · as of 2026-09-20
Input ($/Mtok)$0.15source · as of 2026-09-20 · medium confidence
Output ($/Mtok)$0.47source · as of 2026-09-20 · medium confidence
LicenceProprietarysource · as of 2026-09-20
Max outputNot publishedsource · as of 2026-09-20
ParametersNot publishedsource · as of 2026-09-20
Released18 Sep 2026source · as of 2026-09-20
Model IDqwen3.8-omni-flashsource · as of 2026-09-20 · medium confidence

What matters

  • Takes text, images, audio and video in one pass and returns text
  • Alibaba reports more than 26% average gain over Qwen3.5-Omni-Plus across 30 evaluations
  • No weights published, which is a change of habit for this team

Best for audio and video agents.