Qwen 2.5 Omni 7B

Mar 2025
Qwen 2.5 Omni 7B
Compare in
on
self-host onlyprice
58.9MAEB v1 audio-only
Mar 2025released
33K tokensMax input
enLanguages

Ranking?

Full ranking →
1Larger Clap Music And Speech96.92MuQ MuLan Large96.33Larger Clap General95.64Ast Finetuned Audioset 10 10 0.459394.65Ebind Full93.8
···
27Qwen3 Omni 30B A3B Instruct63.228Qwen 2.5 Omni 7B62.129Speecht5 Multimodal59.6

Alternatives?

ModelScoreCostPriceWhat changes
BetterJina Embeddings v5 Omni NanounstatedOpen+33%—$0.02Max input: 33K tokens → 8K tokens+1lOpenLarger Clap Music And Speech+56%——Max input: not verified+1BetterJina Embeddings v5 Omni SmallunstatedOpen+35%—$0.05Languages: not verifiedOBetterMuQ MuLan LargeOpen+55%——Max input: not verifiedlBetterLarger Clap GeneralOpen+54%——Max input: not verified+1

Run it yourself

estimate
Fits on✓ 8 GB (phone, laptop) · compressed✓ 16 GB · 8-bit✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceApache-2.0free for commercial use

Benchmarks

in Embeddings & Retrieval
MAEB v1 · Audio Zero-shot Classification · English17.7#13
MAEB v1 · Any-to-any Retrieval · multilingual (150 languages)8.6#22
MAEB v1 · Audio Pair Classification · English52.8#20
MAEB v1 · Audio Reranking · No linguistic content58.9#28
MAEB v1 · Audio Multilabel Classification · English9.8#38
MAEB v1 · Audio Clustering · multilingual (6 languages)0.7#40
MAEB v1 audio-only · Any-to-any Retrieval · multilingual (4 languages)62.1#28
MAEB v1 audio-only · Audio Clustering · multilingual (6 languages)0.7#40
MAEB v1 audio-only · Audio Pair Classification · English52.8#20
MAEB v1 audio-only · Audio Reranking · No linguistic content58.9#28
MVEB v1 · Mean over tasks · multilingual (16 languages)12.8#15
MVEB v1 · Mean over task types · multilingual (16 languages)18.5#15
MVEB v1 · Any-to-any Retrieval · English0.5#17
MVEB v1 · Video Zero-shot Classification · English16.5#16
MVEB v1 · Video Pair Classification · English53.2#16
MVEB v1 · Video Classification · multilingual (16 languages)20.2#15
MVEB v1 · Video Clustering · English7.5#15
MVEB v1 · Video Centric QA · English13.2#26
MVEB text video v1 · Mean over tasks · multilingual (16 languages)10.4#23
MVEB text video v1 · Mean over task types · multilingual (16 languages)18.9#23
MVEB text video v1 · Video Zero-shot Classification · English15.5#24
MVEB text video v1 · Any-to-any Retrieval · English0.4#23
MVEB text video v1 · Video Clustering · English7.4#20
MVEB text video v1 · Video Classification · multilingual (16 languages)11.8#24
MVEB text video v1 · Video Centric QA · English25.1#14
MVEB text video v1 · Video Pair Classification · English53.1#27
MVEB video v1 · Mean over tasks · multilingual (16 languages)30.7#24
MVEB video v1 · Mean over task types · multilingual (16 languages)36.6#24
MVEB video v1 · Video Classification · multilingual (16 languages)18.9#24
MVEB video v1 · Video Pair Classification · English54.3#25

FAQ

How much does Qwen 2.5 Omni 7B cost?›

There is no per-unit API price: self-host only.

Is Qwen 2.5 Omni 7B free to use?›

There is no free API. The weights are open, so you can run it yourself.

How good is Qwen 2.5 Omni 7B, and what is it best at?›

#23 of 27 in Multimodal.

Is Qwen 2.5 Omni 7B better than Qwen3 Omni 30B A3B Instruct?›

On MAEB v1 audio-only · Any-to-any Retrieval · multilingual (4 languages), Qwen 2.5 Omni 7B scores 62.1 vs 63.2 — lower.

When was Qwen 2.5 Omni 7B released? Is there a newer version?›

Released 2025-03-22. It is the current version in its line.

Is Qwen 2.5 Omni 7B open source?›

Yes, the weights are open (Apache-2.0 licence).

Can I run Qwen 2.5 Omni 7B locally?›

The weights are open and can be self-hosted; hardware needs are not confirmed yet.

What is Qwen 2.5 Omni 7B?›

Qwen 2.5 Omni 7B is an open-weights embeddings & retrieval model from Alibaba (Qwen), released Mar 2025.

What are the best alternatives to Qwen 2.5 Omni 7B?›

Jina Embeddings v5 Omni Nano (better), Larger Clap Music And Speech (open weights).