Skip to content
Model Menu
Models
Compare
Leaderboard
Benchmarks
⌘K
Top
Chat
Coding
Agents
Writing
Vision
Image
Video
Speech
Embeddings
Decision
Music
Filters
Best for Embeddings & Retrieval
All models
2314
Chat & Knowledge
563
Coding
383
Agents & Tool Use
543
Writing & Language
361
Vision & Documents
397
Image Generation & Editing
184
Video Generation
148
Speech & Voice
358
Embeddings & Retrieval
688
Decision models
118
Music & Sound
29
Embeddings
492
Rerankers
16
Multimodal
167
Audio
13
Video
0
MAEB v1 audio-only (task avg)
MAEB · Audio Classification
MAEB v1 audio-only
All benchmarks
12
Embeddings & Retrieval benchmarks
12
×
Benchmark
By
Leader
Models
MAEB v1 audio-only
Mean over tasks · multilingual (102 languages)
Embeddings classify relationships between audio clips, reorder candidate audio clips by relevance, group similar audio clips, classify audio clips, retrieve matching examples across modalities on Maeb v1 Audio Only · mean over tasks in 102 languages including Afrikaans, Amharic, Armenian, Assamese, Asturian, Belarusian, Bengali, Bosnian; tasks receive equal weight
MTEB
M
Ast Finetuned Audioset 10 10 0.4593
44.7
13
+6 versions & slices
Mean over task types · multilingual (102 languages)
Embeddings classify relationships between audio clips, reorder candidate audio clips by relevance, group similar audio clips, classify audio clips, retrieve matching examples across modalities on Maeb v1 Audio Only · mean over task types in 102 languages including Afrikaans, Amharic, Armenian, Assamese, Asturian, Belarusian, Bengali, Bosnian; task types receive equal weight
MTEB
M
Ast Finetuned Audioset 10 10 0.4593
55.2
13
Any-to-any Retrieval · multilingual (4 languages)
Embeddings retrieve matching examples across modalities on Maeb v1 Audio Only · Any-to-any Retrieval in English, French, German, Spanish
MTEB
M
Ast Finetuned Audioset 10 10 0.4593
94.6
13
Audio Classification · multilingual (102 languages)
Embeddings classify audio clips on Maeb v1 Audio Only · Audio Classification in 102 languages including Afrikaans, Amharic, Armenian, Assamese, Asturian, Belarusian, Bengali, Bosnian
MTEB
M
Ast Finetuned Audioset 10 10 0.4593
45.7
13
Audio Clustering · multilingual (6 languages)
Embeddings group similar audio clips on Maeb v1 Audio Only · Audio Clustering in English, French, German, No linguistic content, Polish, Spanish
MTEB
Vggish
7.8
13
Audio Pair Classification · English
Embeddings classify relationships between audio clips on Maeb v1 Audio Only · Audio Pair Classification in English
MTEB
Yamnet
54.6
13
Audio Reranking · No linguistic content
Embeddings reorder candidate audio clips by relevance on Maeb v1 Audio Only · Audio Reranking in No linguistic content
MTEB
Yamnet
81.7
13
MAEB
v1 · Audio Classification · multilingual (102 languages)
Embeddings classify audio clips on Maeb v1 · Audio Classification in 102 languages including Afrikaans, Amharic, Armenian, Assamese, Asturian, Belarusian, Bengali, Bosnian
MTEB
M
Ast Finetuned Audioset 10 10 0.4593
45.7
13
+4 versions & slices
v1 · Audio Clustering · multilingual (6 languages)
Embeddings group similar audio clips on Maeb v1 · Audio Clustering in English, French, German, No linguistic content, Polish, Spanish
MTEB
Vggish
7.8
13
v1 · Audio Pair Classification · English
Embeddings classify relationships between audio clips on Maeb v1 · Audio Pair Classification in English
MTEB
Yamnet
54.6
13
v1 · Audio Reranking · No linguistic content
Embeddings reorder candidate audio clips by relevance on Maeb v1 · Audio Reranking in No linguistic content
MTEB
Yamnet
81.7
13
v1 · Audio Multilabel Classification · English
Embeddings assign multiple labels to audio clips on Maeb v1 · Audio Multilabel Classification in English
MTEB
M
Ast Finetuned Audioset 10 10 0.4593
72.7
12
No benchmark matches
Model
MAEB v1 audio-only
Dimensions
Languages
Price
Released
M
Ast Finetuned Audioset 10 10 0.4593
Massachusetts Institute of Technology
Open
55.2
768
—
Jul 2021
Vggish
Google
Open
53.5
128
—
Jun 2019
Yamnet
Google
Open
51.1
1,024
—
Oct 2020
m
HeAR S1.1
matthewagi
Open
50.2
384
—
Mar 2026
s
Cnn14 Esc50
speechbrain
Open
44.8
—
en
Nov 2022
Encodec 24khz
Meta
Open
43.4
128
—
Oct 2022
s
M Ctc T Large
speechbrain
Open
39.2
1,536
en
Jan 2022
Unispeech Sat Base 100h Libri Ft
Microsoft
Open
37.8
768
en
Oct 2021
Data2vec Audio Large 960h
Meta
Open
37.8
1,024
en
Feb 2022
a
Sew D Tiny 100K Ft Ls100h
asapp
Open
37
384
en
Sep 2021
Data2vec Audio Base 960h
Meta
Open
36.1
768
en
Feb 2022
a
Sew D Mid 400K Ft Ls100h
asapp
Open
34.1
512
en
Sep 2021
a
Sew D Base Plus 400K Ft Ls100h
asapp
Open
33.7
768
en
Sep 2021
