Qwen/Qwen3-VL-Embedding-8B

Explore Qwen/Qwen3-VL-Embedding-8B: model specifications, capabilities, benchmark results, platform pricing and original sources.

huggingface/Qwen/Qwen3-VL-Embedding-8B

Specifications & capabilities

Published model ID
Qwen3-VL-Embedding-8B

Evaluations

1–6 / 6

MTEB

74.45%

source id
mteb
leaderboard name
MTEB Pair Classification
dimension
embedding
metric
Average task score
rank
64
sample size
7
aggregation scope
splits
  • test
subsets
  • default
task_count
7
task_names
  • AVEDatasetVPairClassification
  • AVSpeakerBenchPairClassification
  • HumanAnimalCartoonVPairClassification
  • MELDVPairClassification
  • MusicAVQAVPairClassification
  • RAVDESSAVVPairClassification
  • VinogroundPairClassification
sample_size
7
subset_count
1
subset_sample
  • default
aggregation_method
mean-of-task-means
model_revision_count
1
model_revision_sample
  • 2c4565515e0f265c6511776e7193b22c0968ddc7
source model id
Qwen/Qwen3-VL-Embedding-8B
Original result ↗

MTEB

51.304%

source id
mteb
leaderboard name
MTEB Clustering
dimension
embedding
metric
Average task score
rank
171
sample size
8
aggregation scope
splits
  • test
subsets
  • default
task_count
8
task_names
  • AVEDatasetVideoClustering
  • HMDB51Clustering
  • MELDEmotionVideoClustering
  • MELDSpeakerVideoClustering
  • MusicAVQACLSVideoClustering
  • RAVDESSVideoClustering
  • UCF101VideoClustering
  • WorldSense1MinDomainVideoClustering
sample_size
8
subset_count
1
subset_sample
  • default
aggregation_method
mean-of-task-means
model_revision_count
1
model_revision_sample
  • 2c4565515e0f265c6511776e7193b22c0968ddc7
source model id
Qwen/Qwen3-VL-Embedding-8B
Original result ↗

MTEB

68.407%

source id
mteb
leaderboard name
MTEB Retrieval
dimension
embedding
metric
Average task score
rank
195
sample size
129
aggregation scope
splits
  • dev
  • test
subsets
  • default
  • eng_Latn-kor_Hang
  • english
  • french
  • german
  • italian
  • ko
  • kor_Hang-eng_Latn
  • kor_Hang-kor_Hang
  • korean
  • portuguese
  • spanish
task_count
46
task_names
  • AVMemeExamT2VRetrieval
  • AVMemeExamV2TRetrieval
  • ActivityNetCaptionsT2VRetrieval
  • ActivityNetCaptionsV2TRetrieval
  • AudioCapsAVT2VRetrieval
  • AudioCapsAVV2TRetrieval
  • AutoRAGRetrieval
  • BelebeleRetrieval
  • DiDeMoT2VRetrieval
  • DiDeMoV2TRetrieval
  • MIRACLRetrieval
  • MSVDT2VRetrieval
  • MSVDV2TRetrieval
  • MrTidyRetrieval
  • MultiLongDocRetrieval
  • Panda70MT2VRetrieval
  • Panda70MV2TRetrieval
  • SQuADKorV1Retrieval
  • Shot2Story20KT2VRetrieval
  • Shot2Story20KV2TRetrieval
  • TUNABenchT2VRetrieval
  • TUNABenchV2TRetrieval
  • VALOR32KT2VRetrieval
  • VALOR32KV2TRetrieval
  • VATEXT2VRetrieval
  • VATEXV2TRetrieval
  • VGGSoundAVT2VRetrieval
  • VGGSoundAVV2TRetrieval
  • Vidore3ComputerScienceRetrieval
  • Vidore3ComputerScienceRetrieval.v2
  • Vidore3EnergyRetrieval
  • Vidore3EnergyRetrieval.v2
  • Vidore3FinanceEnRetrieval
  • Vidore3FinanceEnRetrieval.v2
  • Vidore3FinanceFrRetrieval
  • Vidore3FinanceFrRetrieval.v2
  • Vidore3HrRetrieval
  • Vidore3HrRetrieval.v2
  • Vidore3IndustrialRetrieval
  • Vidore3IndustrialRetrieval.v2
  • Vidore3PharmaceuticalsRetrieval
  • Vidore3PharmaceuticalsRetrieval.v2
  • Vidore3PhysicsRetrieval
  • Vidore3PhysicsRetrieval.v2
  • YouCook2T2VRetrieval
  • YouCook2V2TRetrieval
sample_size
129
subset_count
12
subset_sample
  • default
  • eng_Latn-kor_Hang
  • english
  • french
  • german
  • italian
  • ko
  • kor_Hang-eng_Latn
  • kor_Hang-kor_Hang
  • korean
  • portuguese
  • spanish
aggregation_method
mean-of-task-means
model_revision_count
1
model_revision_sample
  • 2c4565515e0f265c6511776e7193b22c0968ddc7
source model id
Qwen/Qwen3-VL-Embedding-8B
Original result ↗

MTEB

62.766%

source id
mteb
leaderboard name
MTEB Overall
dimension
embedding
metric
Average task score
rank
248
sample size
190
aggregation scope
splits
  • dev
  • test
subsets
  • default
  • eng_Latn-kor_Hang
  • english
  • french
  • german
  • italian
  • ko
  • kor_Hang-eng_Latn
  • kor_Hang-kor_Hang
  • korean
  • portuguese
  • spanish
task_count
107
task_names
  • AVEDatasetVPairClassification
  • AVEDatasetVideoClassification
  • AVEDatasetVideoClustering
  • AVEDatasetVideoZeroShot
  • AVMemeExamT2VRetrieval
  • AVMemeExamV2TRetrieval
  • AVMemeExamVideoCentricQA
  • AVMemeVideoClassification
  • AVMemeVideoZeroShot
  • AVQAVideoCentricQA
  • AVSpeakerBenchPairClassification
  • AVSpeakerBenchVideoCentricQA
  • ActivityNetCaptionsT2VRetrieval
  • ActivityNetCaptionsV2TRetrieval
  • AudioCapsAVT2VRetrieval
  • AudioCapsAVV2TRetrieval
  • AutoRAGRetrieval
  • BelebeleRetrieval
  • BreakfastClassification
  • BreakfastZeroShot
  • DailyOmniVideoCentricQA
  • DiDeMoT2VRetrieval
  • DiDeMoV2TRetrieval
  • Diving48Classification.V2
  • EgoSchemaVideoCentricQA
  • HMDB51Classification
  • HMDB51Clustering
  • HMDB51ZeroShot
  • HumanAnimalCartoonV
  • HumanAnimalCartoonVPairClassification
  • HumanAnimalCartoonZeroShot
  • Kinetics400V
  • Kinetics400ZeroShot
  • Kinetics600V
  • Kinetics600VZeroShot
  • Kinetics700V
  • Kinetics700VZeroShot
  • Ko-StrategyQA
  • LawIRKo
  • MELDEmotionVideoClustering
  • MELDSpeakerVideoClustering
  • MELDVPairClassification
  • MELDVideoClassification
  • MELDVideoZeroShot
  • MIRACLRetrieval
  • MSRVTTT2V
  • MSRVTTV2T
  • MSVDT2VRetrieval
  • MSVDV2TRetrieval
  • MrTidyRetrieval
  • MultiLongDocRetrieval
  • MusicAVQACLSVideoClassification
  • MusicAVQACLSVideoClustering
  • MusicAVQACLSVideoZeroShot
  • MusicAVQAVPairClassification
  • NExTQAVideoCentricQA
  • OmniVideoBenchVideoCentricQA
  • Panda70MT2VRetrieval
  • Panda70MV2TRetrieval
  • PerceptionTestVideoCentricQA
  • PublicHealthQA
  • RAVDESSAVVPairClassification
  • RAVDESSVClassification
  • RAVDESSVZeroShot
  • RAVDESSVideoClustering
  • SQuADKorV1Retrieval
  • Shot2Story20KT2VRetrieval
  • Shot2Story20KV2TRetrieval
  • SomethingSomethingV2Classification
  • TUNABenchT2VRetrieval
  • TUNABenchV2TRetrieval
  • UCF101VideoClassification
  • UCF101VideoClustering
  • UCF101VideoZeroShotClassification
  • VALOR32KT2VRetrieval
  • VALOR32KV2TRetrieval
  • VATEXT2VRetrieval
  • VATEXV2TRetrieval
  • VGGSoundAVT2VRetrieval
  • VGGSoundAVV2TRetrieval
  • VGGSoundV
  • VGGSoundVideoZeroshot
  • VideoMMEShortVideoCentricQA
  • Vidore3ComputerScienceRetrieval
  • Vidore3ComputerScienceRetrieval.v2
  • Vidore3EnergyRetrieval
  • Vidore3EnergyRetrieval.v2
  • Vidore3FinanceEnRetrieval
  • Vidore3FinanceEnRetrieval.v2
  • Vidore3FinanceFrRetrieval
  • Vidore3FinanceFrRetrieval.v2
  • Vidore3HrRetrieval
  • Vidore3HrRetrieval.v2
  • Vidore3IndustrialRetrieval
  • Vidore3IndustrialRetrieval.v2
  • Vidore3PharmaceuticalsRetrieval
  • Vidore3PharmaceuticalsRetrieval.v2
  • Vidore3PhysicsRetrieval
  • Vidore3PhysicsRetrieval.v2
  • VinogroundPairClassification
  • WorldQAVideoCentricQA
  • WorldSense1MinDomainVideoClustering
  • WorldSense1MinVideoCentricQA
  • WorldSenseVideoClassification
  • WorldSenseVideoZeroShot
  • YouCook2T2VRetrieval
  • YouCook2V2TRetrieval
sample_size
190
subset_count
12
subset_sample
  • default
  • eng_Latn-kor_Hang
  • english
  • french
  • german
  • italian
  • ko
  • kor_Hang-eng_Latn
  • kor_Hang-kor_Hang
  • korean
  • portuguese
  • spanish
aggregation_method
mean-of-task-means
model_revision_count
1
model_revision_sample
  • 2c4565515e0f265c6511776e7193b22c0968ddc7
source model id
Qwen/Qwen3-VL-Embedding-8B
Original result ↗

MTEB

58.346%

source id
mteb
leaderboard name
MTEB Other
dimension
embedding
metric
Average task score
rank
282
sample size
34
aggregation scope
splits
  • dev
  • test
subsets
  • default
  • korean
task_count
34
task_names
  • AVEDatasetVideoZeroShot
  • AVMemeExamVideoCentricQA
  • AVMemeVideoZeroShot
  • AVQAVideoCentricQA
  • AVSpeakerBenchVideoCentricQA
  • BreakfastZeroShot
  • DailyOmniVideoCentricQA
  • EgoSchemaVideoCentricQA
  • HMDB51ZeroShot
  • HumanAnimalCartoonV
  • HumanAnimalCartoonZeroShot
  • Kinetics400V
  • Kinetics400ZeroShot
  • Kinetics600V
  • Kinetics600VZeroShot
  • Kinetics700V
  • Kinetics700VZeroShot
  • Ko-StrategyQA
  • LawIRKo
  • MELDVideoZeroShot
  • MSRVTTT2V
  • MSRVTTV2T
  • MusicAVQACLSVideoZeroShot
  • NExTQAVideoCentricQA
  • OmniVideoBenchVideoCentricQA
  • PerceptionTestVideoCentricQA
  • PublicHealthQA
  • RAVDESSVZeroShot
  • VGGSoundV
  • VGGSoundVideoZeroshot
  • VideoMMEShortVideoCentricQA
  • WorldQAVideoCentricQA
  • WorldSense1MinVideoCentricQA
  • WorldSenseVideoZeroShot
sample_size
34
subset_count
2
subset_sample
  • default
  • korean
aggregation_method
mean-of-task-means
model_revision_count
1
model_revision_sample
  • 2c4565515e0f265c6511776e7193b22c0968ddc7
source model id
Qwen/Qwen3-VL-Embedding-8B
Original result ↗

MTEB

54.495%

source id
mteb
leaderboard name
MTEB Classification
dimension
embedding
metric
Average task score
rank
414
sample size
12
aggregation scope
splits
  • test
subsets
  • default
task_count
12
task_names
  • AVEDatasetVideoClassification
  • AVMemeVideoClassification
  • BreakfastClassification
  • Diving48Classification.V2
  • HMDB51Classification
  • MELDVideoClassification
  • MusicAVQACLSVideoClassification
  • RAVDESSVClassification
  • SomethingSomethingV2Classification
  • UCF101VideoClassification
  • UCF101VideoZeroShotClassification
  • WorldSenseVideoClassification
sample_size
12
subset_count
1
subset_sample
  • default
aggregation_method
mean-of-task-means
model_revision_count
1
model_revision_sample
  • 2c4565515e0f265c6511776e7193b22c0968ddc7
source model id
Qwen/Qwen3-VL-Embedding-8B
Original result ↗

Safety assessments

1–1 / 1

Foundation Model Transparency Index 2025 ↗

26 / 100

source id
stanfordFMTI
canonical model id
huggingface/Qwen/Qwen3-VL-Embedding-8B
taxonomy
transparency
hazard category
transparency
hazard name
Organization transparency
policy area
Alibaba
assessment type
organization-level disclosure
metric
disclosure_score
result label
26 / 100
sample size
100
updated at
2026-11-09T18:08:05.227Z

Original sources

1–5 / 5

Hugging Face · Qwen/Qwen3-VL-Embedding-8B ↗

arXiv · Paper · 2601.04720 ↗

GitHub · Catalog ↗

Hugging Face · Catalog ↗

Hugging Face · Catalog ↗

Source links

1–3 / 3

Hugging Face model card

1–1 / 1

Explore more models

Models

Loading
EVIDENCE / COMPARISON

Benchmarks

Explore every benchmark

Loading

These charts contain model-linked results. The charts above also include unlinked source profiles. Compare scores within each evaluation. Model quality, usage, transparency and system performance measure different things.

Install TensorTrace