All tasks / Models
Qwen3-Omni-30B-A3B-Instruct
Alibaba's Apache-2.0 open-weights omni model (text, image, audio, video in; text and speech out), a 30B MoE with 3B active.
- Alibaba (Qwen)
- Open weights
- Released 22 Sep 2025
Where it ranks
- –Voice agentsModels · Local · no result on this task’s benchmarks yet
Facts
- Vendor
- Alibaba (Qwen)
- Family
- Qwen3-Omni
- Released
- 22 Sep 2025
- Weights
- Open: download and run it yourself
- License
- Apache-2.0
- Parameters
- 35.26B (3B active)
- GPU memory to run locally
- ~80 GB (Derived: 35.3B parameters (Hugging Face safetensors count) x 2 bytes = ~70.5 GB of bf16 weights, so an 80 GB-class GPU; the card says disabling audio output saves ~10 GB.)
- API price
- $0.03 per minute of audio (Artificial Analysis 'Cost per Hour of Input Audio' for Qwen3 Omni 30B A3B Instruct (flash) ($1.767/hour: cost of its Big Bench Audio run normalised to input-audio length) divided by 60.)
- Access
- Free (open weights)
- Last verified
- 18 Oct 2026
Price from artificialanalysis.ai, 18 Oct 2026.
Benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| Artificial Analysis Big Bench Audio (speech reasoning) | 58.7% | Qwen3 Omni 30B A3B Instruct (flash)Big Bench Audio speech reasoning (audio in, audio out); embedded value 0.587 as percent; no last-updated date shown. | 18 Oct 2026 | artificialanalysis.ai |
Where you can use it
No tracked tool lists this model yet; as open weights it runs in local runtimes.