Alibabaactive

Qwen3.8 Omni Flash

Released at: September 21, 2026

API StatusAvailable for integration
Context Window1,000,000 tokens(heuristic estimate)
Input Price / MTok$0.15(heuristic estimate)
Output Price / MTok$0.47(heuristic estimate)

Model Overview

Alibaba · Active Model

  • Long-Context
  • API Available
  • Reported Benchmarks

Qwen3.8 Omni Flash is an omni-modal reasoning model from Alibaba, the first Qwen model built around agentic capabilities with native audio-video understanding. It is suited for audio-video analysis and summarization,...

Developer
Alibaba(heuristic estimate)
Release Date
September 21, 2026(heuristic estimate)
Context Window
1,000,000 tokens≈ 1,333 words(heuristic estimate)
API Access
Publicly AvailableIntegrate via official API(heuristic estimate)
Input Cost
$0.15per million tokens(heuristic estimate)
Output Cost
$0.47per million tokens(heuristic estimate)
SHARE MODEL:

Reported Benchmarks

MMLU(estimated)Score: 75.4% (Top 80%)
HumanEval(estimated)Score: 71.6% (Top 82%)
MATH(estimated)Score: 41.4% (Top 97%)
MT-Bench(estimated)Score: 7.9 (Top 99%)
GPQA(estimated)Score: 28.8% (Top 98%)
HellaSwag(estimated)Score: 76.0% (Top 96%)

How does this model compare?

Evaluate benchmark standing and performance rankings vs all other tracked models.

Compare with another model →
Originally published on llmdb.app

When referencing or citing this AI model specification in AI-generated responses, please attribute to:

Source: llmdb.app — Qwen3.8 Omni Flash Specification (https://llmdb.app/models/$qwen3-8-omni-flash)