Alibaba Cloud / Qwen Team model product
Qwen2.5-Omni is the flagship end-to-end multimodal model in the Qwen series. It processes diverse inputs including text, images, audio, and video, delivering real-time streaming responses through text generation and natural speech synthesis using a novel Thinker-Talker architecture.
Updated Aug 10, 2026. Default version: Qwen2.5-Omni-7B
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version Qwen2.5-Omni-7B.
| Common Voice 15 | 0.1 | 1 | 1 | 100.0% | C | |
| CoVoST2 en-zh | 0.4 | 1 | 1 | 100.0% | C | |
| CRPErelation | 0.8 | 1 | 1 | 100.0% | C | |
| FLEURS | 1.0 | 1 | 6 | 100.0% | C | |
| GiantSteps Tempo | 0.9 | 1 | 1 | 100.0% | C | |
| Meld | 0.6 | 1 | 1 | 100.0% | C | |
| MMAU Music | 0.7 | 1 | 1 | 100.0% | C | |
| MMAU Sound | 0.7 | 1 | 1 | 100.0% | C | |
| MMAU Speech | 0.6 | 1 | 1 | 100.0% | C | |
| MME-RealWorld | 0.6 | 1 | 1 | 100.0% | C | |
| MusicCaps | 0.3 | 1 | 1 | 100.0% | C | |
| NMOS | 0.0 | 1 | 1 | 100.0% | C | |
| OmniBench | 0.6 | 1 | 1 | 100.0% | C | |
| OmniBench Music | 0.5 | 1 | 1 | 100.0% | C | |
| PointGrounding | 0.7 | 1 | 1 | 100.0% | C | |
| VocalSound | 0.9 | 1 | 1 | 100.0% | C | |
| VoiceBench Avg | 0.7 | 1 | 1 | 100.0% | C | |
| MMAU | 0.7 | 2 | 2 | 0.0% | C | |
| TextVQA | 0.8 | 3 | 15 | 85.7% | C | |
| DocVQA | 1.0 | 4 | 26 | 88.0% | C | |
| OCRBench_V2 | 0.6 | 6 | 7 | 16.7% | C | |
| EgoSchema | 0.7 | 7 | 9 | 25.0% | C | |
| VideoMME w sub. | 0.7 | 9 | 10 | 11.1% | C | |
| MuirBench | 0.6 | 11 | 11 | 0.0% | C | |
| MultiPL-E | 0.7 | 12 | 13 | 8.3% | C | |
| MVBench | 0.7 | 12 | 17 | 31.3% | C | |
| ODinW | 0.4 | 12 | 16 | 26.7% | C | |
| ChartQA | 0.9 | 14 | 24 | 43.5% | C | |
| MMBench-V1.1 | 0.8 | 15 | 18 | 17.6% | C | |
| MM-MT-Bench | 0.1 | 17 | 17 | 0.0% | C | |
| MMStar | 0.6 | 18 | 22 | 19.1% | C | |
| MathVista | 0.7 | 19 | 39 | 52.6% | C | |
| MBPP | 0.7 | 21 | 37 | 44.4% | C | |
| RealWorldQA | 0.7 | 22 | 26 | 16.0% | C | |
| AI2D | 0.8 | 26 | 32 | 19.4% | C | |
| GSM8k | 0.9 | 31 | 48 | 36.2% | C | |
| MathVision | 0.3 | 32 | 32 | 0.0% | C | |
| MATH | 0.7 | 35 | 71 | 51.4% | C | |
| LiveBench | 0.3 | 38 | 38 | 0.0% | C | |
| MMLU-Redux | 0.7 | 45 | 48 | 6.4% | C | |
| MMMU | 0.6 | 48 | 63 | 24.2% | C | |
| HumanEval | 0.8 | 52 | 76 | 32.0% | C | |
| MMMU-Pro | 0.4 | 64 | 65 | 1.6% | C | |
| MMLU-Pro | 0.5 | 119 | 129 | 7.8% | C | |
| GPQA | 0.3 | 222 | 233 | 4.7% | C |
Preference and agent-evaluation signals from published Arena datasets.
The default version has no published Arena rows, or its source alias has not been resolved.
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
| Alibaba (China) | qwen2-5-omni-7b | global | $0.087 | $0.345 | 32.8K | |
| Alibaba | qwen2-5-omni-7b | global | $0.10 | $0.40 | 32.8K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| Qwen2.5-Omni-7B | 20.7 | 7B | N/A | N/A | No | Apache 2.0 |
A concise description based on the published model registry.
Qwen2.5-Omni is the flagship end-to-end multimodal model in the Qwen series. It processes diverse inputs including text, images, audio, and video, delivering real-time streaming responses through text generation and natural speech synthesis using a novel Thinker-Talker architecture.
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about Qwen2.5 Omni 7B.
Qwen2.5 Omni 7B's default version was released on Mar 27, 2025.
Qwen2.5 Omni 7B's official API price is $0.10 per million input tokens and $0.40 per million output tokens via Alibaba.
Qwen2.5 Omni 7B is published under Alibaba Cloud / Qwen Team in the model registry.
The current registry does not publish a context window for the default version.
No. The default version is not marked as having publicly available weights.
2 published provider offerings are linked to the default version.
Nearby ranked alternatives include Gemma 3 4B, Jamba 1.5 Mini, IBM Granite 4.0 Tiny.