OpenAI: Text Embedding 3 Small
text-embedding-3-small is OpenAI's improved, more performant version of the ada embedding model. Embeddings are a numerical representation of text that can be used to measure the relatedness between …
openai/text-embedding-3-small
text-embedding-3-small is OpenAI's improved, more performant version of the ada embedding model. Embeddings are a numerical representation of text that can be used to measure the relatedness between two pieces of text. Embeddings are useful for search, clustering, recommendations, anomaly detection, and classification tasks.
Specifications & capabilities
- Input
- text
- Output
- embeddings
- Context window · tokens
- 8192
- Capabilities
- embeddings
Description source ↗Evaluations
1–20 / 33
90.924%
- source id
- mteb
- leaderboard name
- MTEB Pair Classification
- dimension
- embedding
- metric
- Average task score
- rank
- 9
- sample size
- 12
- aggregation scope
- task_count
- 1
- task_names
- PubChemWikiPairClassification
- sample_size
- 12
- subset_count
- 12
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗69.34%
- source id
- mteb
- leaderboard name
- MTEB Summarization
- dimension
- embedding
- metric
- Average task score
- rank
- 28
- sample size
- 1
- aggregation scope
- task_count
- 1
- sample_size
- 1
- subset_count
- 1
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗66.86%
- source id
- mteb
- leaderboard name
- MTEB Bitext Mining
- dimension
- embedding
- metric
- Average task score
- rank
- 60
- sample size
- 2
- aggregation scope
- task_count
- 2
- task_names
- BornholmBitextMining
- NorwegianCourtsBitextMining
- sample_size
- 2
- subset_count
- 1
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗74.995%
- source id
- mteb
- leaderboard name
- MTEB Reranking
- dimension
- embedding
- metric
- Average task score
- rank
- 67
- sample size
- 2
- aggregation scope
- task_count
- 2
- task_names
- CMedQAv2-reranking
- SyntecReranking
- sample_size
- 2
- subset_count
- 1
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗65.384%
- source id
- mteb
- leaderboard name
- MTEB Bitext Mining
- dimension
- embedding
- metric
- Average task score
- rank
- 73
- sample size
- 2
- aggregation scope
- task_count
- 2
- task_names
- BornholmBitextMining
- NorwegianCourtsBitextMining
- sample_size
- 2
- subset_count
- 1
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small (embed_dim=512)
Original result ↗41.522%
- source id
- mteb
- leaderboard name
- MTEB Bitext Mining
- dimension
- embedding
- metric
- Average task score
- rank
- 146
- sample size
- 45634
- aggregation scope
- splits
- devtest
- test
- train
- validation
- task_count
- 9
- task_names
- BibleNLPBitextMining
- DiaBlaBitextMining
- FloresBitextMining
- IN22GenBitextMining
- IndicGenBenchFloresBitextMining
- NTREXBitextMining
- NollySentiBitextMining
- NusaTranslationBitextMining
- NusaXBitextMining
- sample_size
- 45634
- subset_count
- 43708
- subset_sample
- aai_Latn-eng_Latn
- aak_Arab-eng_Latn
- aau_Latn-eng_Latn
- aaz_Latn-eng_Latn
- abt_Latn-eng_Latn
- abx_Latn-eng_Latn
- aby_Latn-eng_Latn
- ace_Arab-ace_Latn
- ace_Arab-acm_Arab
- ace_Arab-acq_Arab
- ace_Arab-aeb_Arab
- ace_Arab-afr_Latn
- ace_Arab-ajp_Arab
- ace_Arab-aka_Latn
- ace_Arab-als_Latn
- ace_Arab-amh_Ethi
- ace_Arab-apc_Arab
- ace_Arab-arb_Arab
- ace_Arab-arb_Latn
- ace_Arab-ars_Arab
- ace_Arab-ary_Arab
- ace_Arab-arz_Arab
- ace_Arab-asm_Beng
- ace_Arab-ast_Latn
- ace_Arab-awa_Deva
- ace_Arab-ayr_Latn
- ace_Arab-azb_Arab
- ace_Arab-azj_Latn
- ace_Arab-bak_Cyrl
- ace_Arab-bam_Latn
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗29.152%
- source id
- mteb
- leaderboard name
- MTEB Bitext Mining
- dimension
- embedding
- metric
- Average task score
- rank
- 176
- sample size
- 512
- aggregation scope
- task_count
- 4
- task_names
- IN22ConvBitextMining
- LinceMTBitextMining
- PhincBitextMining
- PubChemSMILESBitextMining
- sample_size
- 512
- subset_count
- 511
- subset_sample
- asm_Beng-ben_Beng
- asm_Beng-brx_Deva
- asm_Beng-doi_Deva
- asm_Beng-eng_Latn
- asm_Beng-gom_Deva
- asm_Beng-guj_Gujr
- asm_Beng-hin_Deva
- asm_Beng-kan_Knda
- asm_Beng-kas_Arab
- asm_Beng-mai_Deva
- asm_Beng-mal_Mlym
- asm_Beng-mar_Deva
- asm_Beng-mni_Mtei
- asm_Beng-npi_Deva
- asm_Beng-ory_Orya
- asm_Beng-pan_Guru
- asm_Beng-san_Deva
- asm_Beng-sat_Olck
- asm_Beng-snd_Deva
- asm_Beng-tam_Taml
- asm_Beng-tel_Telu
- asm_Beng-urd_Arab
- ben_Beng-asm_Beng
- ben_Beng-brx_Deva
- ben_Beng-doi_Deva
- ben_Beng-eng_Latn
- ben_Beng-gom_Deva
- ben_Beng-guj_Gujr
- ben_Beng-hin_Deva
- ben_Beng-kan_Knda
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗63.972%
- source id
- mteb
- leaderboard name
- MTEB Reranking
- dimension
- embedding
- metric
- Average task score
- rank
- 180
- sample size
- 3
- aggregation scope
- task_count
- 2
- task_names
- MIRACLReranking
- WikipediaRerankingMultilingual
- sample_size
- 3
- subset_count
- 2
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗70.068%
- source id
- mteb
- leaderboard name
- MTEB Classification
- dimension
- embedding
- metric
- Average task score
- rank
- 184
- sample size
- 39
- aggregation scope
- subsets
- arabic
- default
- english
- french
- german
- hindi
- italian
- portuguese
- spanish
- task_count
- 31
- task_names
- GeoreviewClassification
- HeadlineClassification
- HindiDiscourseClassification
- InappropriatenessClassification
- KinopoiskClassification
- MalayalamNewsClassification
- RuReviewsClassification
- RuSciBenchGRNTIClassification
- RuSciBenchOECDClassification
- SDSEyeProtectionClassification
- SDSGlovesClassification
- SanskritShlokasClassification
- SensitiveTopicsClassification
- ToxicChatClassification
- TweetSentimentClassification
- UrduRomanSentimentClassification
- WikipediaBioMetChemClassification
- WikipediaBiolumNeurochemClassification
- WikipediaChemEngSpecialtiesClassification
- WikipediaChemFieldsClassification
- WikipediaChemistryTopicsClassification
- WikipediaCompChemSpectroscopyClassification
- WikipediaCryobiologySeparationClassification
- WikipediaCrystallographyAnalyticalClassification
- WikipediaGreenhouseEnantiopureClassification
- WikipediaIsotopesFissionClassification
- WikipediaLuminescenceClassification
- WikipediaOrganicInorganicClassification
- WikipediaSaltsSemiconductorsClassification
- WikipediaSolidStateColloidalClassification
- WikipediaTheoreticalAppliedClassification
- sample_size
- 39
- subset_count
- 9
- subset_sample
- arabic
- default
- english
- french
- german
- hindi
- italian
- portuguese
- spanish
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗77.626%
- source id
- mteb
- leaderboard name
- MTEB Semantic Similarity
- dimension
- embedding
- metric
- Average task score
- rank
- 211
- sample size
- 7
- aggregation scope
- subsets
- default
- en
- en-en
- es
- es-en
- es-es
- task_count
- 4
- sample_size
- 7
- subset_count
- 6
- subset_sample
- default
- en
- en-en
- es
- es-en
- es-es
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗48.029%
- source id
- mteb
- leaderboard name
- MTEB Clustering
- dimension
- embedding
- metric
- Average task score
- rank
- 213
- sample size
- 14
- aggregation scope
- splits
- articles
- headlines
- summaries
- test
- validation
- task_count
- 11
- task_names
- BeytooteClustering
- DigikalamagClustering
- MLSUMClusteringP2P
- MLSUMClusteringS2S
- NLPTwitterAnalysisClustering
- SNLHierarchicalClusteringP2P
- SNLHierarchicalClusteringS2S
- SwednClusteringP2P
- SwednClusteringS2S
- VGHierarchicalClusteringP2P
- VGHierarchicalClusteringS2S
- sample_size
- 14
- subset_count
- 2
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗57.489%
- source id
- mteb
- leaderboard name
- MTEB Pair Classification
- dimension
- embedding
- metric
- Average task score
- rank
- 228
- sample size
- 14
- aggregation scope
- task_count
- 1
- sample_size
- 14
- subset_count
- 7
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗46.691%
- source id
- mteb
- leaderboard name
- MTEB Clustering
- dimension
- embedding
- metric
- Average task score
- rank
- 239
- sample size
- 250
- aggregation scope
- task_count
- 26
- task_names
- AlloProfClusteringS2S.v2
- ArXivHierarchicalClusteringP2P
- ArXivHierarchicalClusteringS2S
- ArxivClusteringP2P
- ArxivClusteringS2S
- BigPatentClustering.v2
- BiorxivClusteringP2P
- BiorxivClusteringP2P.v2
- BiorxivClusteringS2S
- CLSClusteringP2P.v2
- HALClusteringS2S.v2
- MasakhaNEWSClusteringS2S
- MedrxivClusteringP2P
- MedrxivClusteringP2P.v2
- MedrxivClusteringS2S
- PlscClusteringP2P.v2
- RedditClustering
- RedditClusteringP2P
- RomaniBibleClustering
- SIB200ClusteringS2S
- StackExchangeClustering
- StackExchangeClustering.v2
- StackExchangeClusteringP2P
- TwentyNewsgroupsClustering
- WikiCitiesClustering
- WikiClusteringP2P.v2
- sample_size
- 250
- subset_count
- 228
- subset_sample
- ace_Latn
- acm_Arab
- acq_Arab
- aeb_Arab
- afr_Latn
- ajp_Arab
- aka_Latn
- als_Latn
- amh
- amh_Ethi
- apc_Arab
- arb_Latn
- ars_Arab
- ary_Arab
- arz_Arab
- asm_Beng
- ast_Latn
- awa_Deva
- ayr_Latn
- azb_Arab
- azj_Latn
- bak_Cyrl
- bam_Latn
- ban_Latn
- bel_Cyrl
- bem_Latn
- ben_Beng
- bho_Deva
- bjn_Latn
- bod_Tibt
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗32.492%
- source id
- mteb
- leaderboard name
- MTEB Summarization
- dimension
- embedding
- metric
- Average task score
- rank
- 247
- sample size
- 1
- aggregation scope
- task_count
- 1
- sample_size
- 1
- subset_count
- 1
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗74.795%
- source id
- mteb
- leaderboard name
- MTEB Semantic Similarity
- dimension
- embedding
- metric
- Average task score
- rank
- 252
- sample size
- 23
- aggregation scope
- subsets
- de
- default
- en
- es
- fr
- it
- nl
- pl
- pt
- ru
- zh
- task_count
- 4
- task_names
- JSTS
- RUParaPhraserSTS
- RuSTSBenchmarkSTS
- STSBenchmarkMultilingualSTS
- sample_size
- 23
- subset_count
- 11
- subset_sample
- de
- default
- en
- es
- fr
- it
- nl
- pl
- pt
- ru
- zh
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗64.982%
- source id
- mteb
- leaderboard name
- MTEB Retrieval
- dimension
- embedding
- metric
- Average task score
- rank
- 259
- sample size
- 107
- source model id
- openai/text-embedding-3-small
Original result ↗63.982%
- source id
- mteb
- leaderboard name
- MTEB Classification
- dimension
- embedding
- metric
- Average task score
- rank
- 259
- sample size
- 236
- source model id
- openai/text-embedding-3-small
Original result ↗54.343%
- source id
- mteb
- leaderboard name
- MTEB Reranking
- dimension
- embedding
- metric
- Average task score
- rank
- 262
- sample size
- 10
- aggregation scope
- splits
- dev
- test
- test_cat
- test_geo
- test_iid
- test_vis
- test_web
- validation
- task_count
- 5
- task_names
- AlloprofReranking
- RuBQReranking
- T2Reranking
- VoyageMMarcoReranking
- WebLINXCandidatesReranking
- sample_size
- 10
- subset_count
- 1
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗62.526%
- source id
- mteb
- leaderboard name
- MTEB Retrieval
- dimension
- embedding
- metric
- Average task score
- rank
- 288
- sample size
- 5
- aggregation scope
- task_count
- 5
- task_names
- DanFeverRetrieval
- NorQuadRetrieval
- SNLRetrieval
- TV2Nordretrieval
- TwitterHjerneRetrieval
- sample_size
- 5
- subset_count
- 1
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small (embed_dim=512)
Original result ↗57.744%
- source id
- mteb
- leaderboard name
- MTEB Other
- dimension
- embedding
- metric
- Average task score
- rank
- 294
- sample size
- 73
- aggregation scope
- splits
- all
- dentistry_and_oral_health
- dermatology
- gastroenterology
- genetics
- neuroscience_and_neurology
- orthopedic_surgery
- otorhinolaryngology
- plastic_surgery
- psychiatry_and_psychology
- pulmonology
- test
- train
- validation
- subsets
- arabic
- chinese
- default
- en
- english
- es
- fr
- french
- korean
- russian
- spanish
- vietnamese
- task_count
- 33
- task_names
- AllegroReviews
- BengaliSentimentAnalysis
- CBD
- CDSC-E
- CDSC-R
- CUREv1
- ClimateFEVERHardNegatives
- CodeFeedbackMT
- CodeFeedbackST
- CodeTransOceanContest
- CodeTransOceanDL
- CosQA
- FEVERHardNegatives
- FalseFriendsGermanEnglish
- GerDaLIRSmall
- HotpotQAHardNegatives
- LeCaRDv2
- LegalBenchConsumerContractsQA
- PSC
- PolEmo2.0-IN
- PubChemAISentenceParaphrasePC
- PubChemSMILESPC
- PubChemSynonymPC
- PubChemWikiParagraphsPC
- PublicHealthQA
- SICK-E-PL
- SICK-R-PL
- SICKFr
- SciFact-PL
- SyntheticText2SQL
- TRECCOVID-PL
- TempReasonL2Context
- TempReasonL3Context
- sample_size
- 73
- subset_count
- 12
- subset_sample
- arabic
- chinese
- default
- en
- english
- es
- fr
- french
- korean
- russian
- spanish
- vietnamese
- aggregation_method
- mean-of-task-means
- model_revision_count
- 1
- source model id
- openai/text-embedding-3-small
Original result ↗Published prices
1–11 / 11
$0.02
- Inference provider
- openRouter
- Platform
- OpenRouter
- Unit
- input_tokens_per_million
- Currency
- USD
$0.00
- Inference provider
- openRouter
- Platform
- OpenRouter
- Unit
- output_tokens_per_million
- Currency
- USD
$0.01
- Inference provider
- openRouter
- Platform
- OpenRouter
- Unit
- input_tokens_per_million
- Currency
- USD
- Billing conditions
- mode
- batch
$0.00
- Inference provider
- openRouter
- Platform
- OpenRouter
- Unit
- output_tokens_per_million
- Currency
- USD
- Billing conditions
- mode
- batch
$0.02
- Inference provider
- vercelAIGatewayCatalog
- Platform
- azure
- Unit
- input_tokens_per_million
- Currency
- USD
- Billing conditions
- endpoint
- azure | openai/text-embedding-3-small
- model_type
- embedding
- billing_scope
- via Vercel AI Gateway
- Billing label
- Input tokens / 1M
- Pricing via
- Vercel AI Gateway
$0.02
- Inference provider
- vercelAIGatewayCatalog
- Platform
- openai
- Unit
- input_tokens_per_million
- Currency
- USD
- Billing conditions
- endpoint
- openai | openai/text-embedding-3-small
- model_type
- embedding
- billing_scope
- via Vercel AI Gateway
- Billing label
- Input tokens / 1M
- Pricing via
- Vercel AI Gateway
$0.02
- Inference provider
- vercelAIGatewayCatalog
- Platform
- openai
- Unit
- input_tokens_per_million
- Currency
- USD
- Billing conditions
- region
- scope
- zone
- geo_region
- eu
- endpoint
- openai | openai/text-embedding-3-small
- model_type
- embedding
- billing_scope
- via Vercel AI Gateway
- Billing label
- Input tokens / 1M
- Pricing via
- Vercel AI Gateway
$0.02
- Inference provider
- vercelAIGatewayCatalog
- Platform
- openai
- Unit
- input_tokens_per_million
- Currency
- USD
- Billing conditions
- region
- scope
- zone
- geo_region
- us
- endpoint
- openai | openai/text-embedding-3-small
- model_type
- embedding
- billing_scope
- via Vercel AI Gateway
- Billing label
- Input tokens / 1M
- Pricing via
- Vercel AI Gateway
$0.02
- Inference provider
- vercelAIGatewayCatalog
- Platform
- Vercel AI Gateway
- Unit
- input_tokens_per_million
- Currency
- USD
- Billing conditions
- model_type
- embedding
- quote_scope
- gateway_catalog
- billing_scope
- Vercel public catalog; endpoint prices may differ
- Billing label
- Input tokens / 1M
- Pricing via
- Vercel AI Gateway
$0.02
- Inference provider
- vercelAIGatewayCatalog
- Platform
- Vercel AI Gateway
- Unit
- input_tokens_per_million
- Currency
- USD
- Billing conditions
- region
- eu
- model_type
- embedding
- quote_scope
- gateway_catalog
- billing_scope
- Vercel public catalog; endpoint prices may differ
- Billing label
- Input tokens / 1M
- Pricing via
- Vercel AI Gateway
$0.02
- Inference provider
- vercelAIGatewayCatalog
- Platform
- Vercel AI Gateway
- Unit
- input_tokens_per_million
- Currency
- USD
- Billing conditions
- region
- us
- model_type
- embedding
- quote_scope
- gateway_catalog
- billing_scope
- Vercel public catalog; endpoint prices may differ
- Billing label
- Input tokens / 1M
- Pricing via
- Vercel AI Gateway
Provider specifications & access
1–2 / 2
- source model id
- openai/text-embedding-3-small
- hosting
- vercel-gateway
- source model id
- text-embedding-3-small
- hosting
- gateway
Safety assessments
1–1 / 1
35 / 100
- source id
- stanfordFMTI
- canonical model id
- openai/text-embedding-3-small
- taxonomy
- transparency
- hazard category
- transparency
- hazard name
- Organization transparency
- policy area
- OpenAI
- assessment type
- organization-level disclosure
- metric
- disclosure_score
- result label
- 35 / 100
- sample size
- 100
- updated at
- 2026-11-09T18:08:05.227Z
Source specification observations
1–4 / 4
- field
- context_length
- value
- 8192
- normalized value
- 8192
- scope
- source-model
- source id
- openRouter
- confidence
- 1
- field
- context_length
- value
- 8192
- normalized value
- 8192
- scope
- source-model
- source id
- openRouter
- confidence
- 1
- field
- max_output_tokens
- value
- 7372
- normalized value
- 7372
- scope
- source-model
- source id
- openRouter
- confidence
- 1
- field
- max_output_tokens
- value
- 7372
- normalized value
- 7372
- scope
- source-model
- source id
- openRouter
- confidence
- 1
Explore more models