Tatoeba sentence retrieval
Models find the English translation of a sentence among competing sentences using multilingual embeddings.
The LASER evaluation uses parallel sentences from Tatoeba.
- Made by
- Facebook AI Research
- Released
- December 2018
- Kind
- Translation
- Level
- School
- Scored
- Checked automatically
- Answers
- Prediction
- Tasks
- All public
- Measures
- Capability
- Subject
- Languages and literature
- Language
- Many languages
- Interaction
- Single turn
- Items from
- Another benchmark
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Many languages
- Last model added
- October 2026
- Scores refreshed
- November 20th, 2026
Gemini Embedding 2 leads with 99.5% among models released by October 2026 Gemini Embedding 2 leads with 99.5% among models released by October 2026
Best score per entry from MTEB’s leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in October 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6.1 Terra
- and 16 more
- 99.5
Gemini Embedding 2
- 87.7 microsoft/harrier-oss-v1-27b
- 85.6 tencent/KaLM-Embedding-Gemma3-12B-2511
- 83.7 intfloat/multilingual-e5-large-instruct
- 82.8 codefuse-ai/F2LLM-v2-14B
- 82.1 facebook/SONAR (Sentence-Level Multimodal and Language-Agnostic Representations)
- 82.0 google/gemini-embedding-001
- 81.5 nvidia/llama-embed-nemotron-8b
- 81.1 sentence-transformers/LaBSE
- 81.0 codefuse-ai/F2LLM-v2-8B
- 80.5 OrdalieTech/Solon-embeddings-large-0.1
- 79.7
Qwen3-Embedding-8B
- 79.5 microsoft/harrier-oss-v1-0.6b
- 78.9 Octen/Octen-Embedding-8B
- 78.4 microsoft/harrier-oss-v1-270m
- 77.7 codefuse-ai/F2LLM-v2-4B
- 76.3 Omartificial-Intelligence-Space/Arabic-labse-Matryoshka
- 76.3
Qwen3-Embedding-4B
- 76.1 Lajavaness/bilingual-embedding-large
- 76.0 Bytedance/Seed1.6-embedding-1215
- 75.2 GritLM/GritLM-8x7B
- 74.1 hotchpotch/bekko-embedding-v1-a25m
- 73.8 BAAI/bge-m3
- 73.8 BidirLM/BidirLM-1B-Embedding
- 73.7 codefuse-ai/F2LLM-v2-1.7B
- 73.5 Alibaba-NLP/gte-Qwen2-7B-instruct
- 72.8 BidirLM/BidirLM-Omni-2.5B-Embedding
- 72.5 Cohere/Cohere-embed-multilingual-v3.0
- 72.3 manu/bge-m3-custom-fr
- 72.3 ai-sage/Giga-Embeddings-instruct-10B-A1.8B-0826
- 72.1 GritLM/GritLM-7B
- 71.3 Mira190/Euler-Legal-Embedding-V1
- 71.3 Linq-AI-Research/Linq-Embed-Mistral
- 71.1 jinaai/jina-embeddings-v3
- 71.1 BidirLM/BidirLM-1.7B-Embedding
- 70.5 Salesforce/SFR-Embedding-Mistral
- 70.2 google/text-multilingual-embedding-002
- 70.1 intfloat/e5-mistral-7b-instruct
- 69.3 Lajavaness/bilingual-embedding-base
- 68.9 intfloat/multilingual-e5-small
- 68.1 intfloat/multilingual-e5-base
- 68.0 ICT-TIME-and-Querit/BOOM_4B_v1
- 68.0 ICT-TIME-and-Querit/ICT-TIME-and-Querit-embedding-v1
- 67.8 mongodb/voyage-multilingual-2
- 67.6 hotchpotch/bekko-embedding-v1-a8m
- 67.2 Lajavaness/bilingual-embedding-small
- 67.2 ai-sage/Giga-Embeddings-instruct-3B-0826
- 66.8 Snowflake/snowflake-arctic-embed-l-v2.0
- 66.0 Salesforce/SFR-Embedding-2_R
- 65.6 openai/text-embedding-3-large
- 65.6 codefuse-ai/F2LLM-v2-0.6B
- 64.3 omarelshehy/arabic-english-sts-matryoshka
- 63.6 Omartificial-Intelligence-Space/Arabic-all-nli-triplet-Matryoshka
- 63.0 nvidia/NV-Embed-v2
- 62.4 codefuse-ai/F2LLM-v2-330M
- 62.2 mongodb/voyage-3
- 62.2 sentence-transformers/paraphrase-multilingual-mpnet-base-v2
- 61.9 mongodb/voyage-3-lite
- 61.1 jinaai/jina-embeddings-v5-omni-small
- 61.1 jinaai/jina-embeddings-v5-text-small
- 60.4 BidirLM/BidirLM-0.6B-Embedding
- 59.5 Gameselo/STS-multilingual-mpnet-base-v2
- 59.3 mongodb/voyage-3.5
- 58.6 Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet
- 58.0
Qwen3-Embedding-0.6B
- 56.9 ibm-granite/granite-embedding-278m-multilingual
- 56.8 Alibaba-NLP/gte-Qwen1.5-7B-instruct
- 56.6 sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
- 56.5 jinaai/jina-embeddings-v5-omni-nano
- 56.5 jinaai/jina-embeddings-v5-text-nano
- 55.7 ibm-granite/granite-embedding-311m-multilingual-r2
- 54.3 nomic-ai/nomic-embed-text-v2-moe
- 54.2
EmbeddingGemma 2
- 54.0 Alibaba-NLP/gte-Qwen2-1.5B-instruct
- 53.7 shibing624/text2vec-base-multilingual
- 53.4 BidirLM/BidirLM-270M-Embedding
- 52.7 Cohere/Cohere-embed-multilingual-light-v3.0
- 51.8 codefuse-ai/F2LLM-v2-160M
- 51.4 google/embeddinggemma-300m
- 50.6 HIT-TMG/KaLM-embedding-multilingual-mini-v1
- 50.2 sentence-transformers/static-similarity-mrl-multilingual-v1
- 50.1 HIT-TMG/KaLM-embedding-multilingual-mini-instruct-v1
- 49.3 ibm-granite/granite-embedding-107m-multilingual
- 48.9 openai/text-embedding-3-small
- 47.5 NovaSearch/jasper_en_vision_language_v1
- 46.5 zeroentropy/zembed-1
- 46.2 codefuse-ai/F2LLM-v2-80M
- 45.7 nvidia/NV-Embed-v1
- 44.2 NovaSearch/stella_en_1.5B_v5
- 42.8 mongodb/voyage-large-2-instruct
- 40.8 Snowflake/snowflake-arctic-embed-m-v2.0
- 39.8 hotchpotch/bekko-static-embedding-v1
- 38.8 ibm-granite/granite-embedding-97m-multilingual-r2
- 38.3 amazon/Titan-text-embeddings-v2
- 35.2 izhx/udever-bloom-7b1
- 32.6 minishlab/potion-multilingual-128M
- 32.5 izhx/udever-bloom-3b
- 29.6 bigscience/sgpt-bloom-7b1-msmarco
- 28.6 izhx/udever-bloom-1b1
- 27.6 cnmoro/static-nomic-384-pten-v2-st
- 23.1 deepfile/embedder-100p
- 23.1 izhx/udever-bloom-560m
- 21.0 sentence-transformers/sentence-t5-xl
- 20.0 NovaSearch/stella_en_400M_v5
- 18.5 sentence-transformers/gtr-t5-xl
- 18.0 manu/sentence_croissant_alpha_v0.2
- 17.8 VPLabs/SearchMap_Preview
- 17.4 manu/sentence_croissant_alpha_v0.3
- 16.7 manu/sentence_croissant_alpha_v0.4
- 16.3 sentence-transformers/sentence-t5-large
- 11.7 intfloat/e5-large-v2
- 11.3 Cohere/Cohere-embed-english-v3.0
- 11.0 Haon-Chen/speed-embedding-7b-instruct
- 9.79 intfloat/e5-large
- 9.06 intfloat/e5-base-v2
- 8.76 sdadas/mmlw-e5-large
- 8.17 aari1995/German_Semantic_STS_V2
- 8.11 intfloat/e5-base
- 8.04 WhereIsAI/UAE-Large-V1
- 7.93 ibm-granite/granite-embedding-125m-english
- 7.89 sdadas/mmlw-roberta-large
- 7.74 intfloat/e5-small
- 7.59 mixedbread-ai/mxbai-embed-large-v1
- 7.57 ai-forever/ru-en-RoSBERTa
- 7.51 BAAI/bge-large-en-v1.5
- 7.49 nomic-ai/nomic-embed-text-v1-unsupervised
- 7.31 intfloat/e5-small-v2
- 7.21 avsolatorio/GIST-large-Embedding-v0
- 7.18 jinaai/jina-embedding-b-en-v1
- 7.12 sdadas/mmlw-roberta-base
- 7.09 sdadas/mmlw-e5-small
- 7.02 thenlper/gte-large
- 6.89 nomic-ai/nomic-embed-text-v1-ablated
- 6.87 sdadas/mmlw-e5-base
- 6.71 infgrad/stella-base-en-v2
- 6.57 dwzhu/e5-base-4k
- 6.39 avsolatorio/GIST-Embedding-v0
- 6.38 thenlper/gte-base
- 6.34 BAAI/bge-base-en-v1.5
- 6.14 ibm-granite/granite-embedding-30m-english
- 6.10 silma-ai/silma-embeddding-matryoshka-v0.1
- 5.89 Cohere/Cohere-embed-english-light-v3.0
- 5.83 nomic-ai/nomic-embed-text-v1.5
- 5.83 nomic-ai/nomic-embed-text-v1
- 5.74 jinaai/jina-embedding-s-en-v1
- 5.49 BAAI/bge-small-en-v1.5
- 5.45 avsolatorio/NoInstruct-small-Embedding-v0
- 5.45 cointegrated/rubert-tiny2
- 5.37 thenlper/gte-small
- 5.31 avsolatorio/GIST-small-Embedding-v0
- 5.21 abhinand/MedEmbed-small-v0.1
- 5.10 sentence-transformers/all-MiniLM-L12-v2
- 4.66 avsolatorio/GIST-all-MiniLM-L6-v2
- 4.64 Mihaiii/Ivysaur
- 4.55 Omartificial-Intelligence-Space/Arabert-all-nli-triplet-Matryoshka
- 4.40 sentence-transformers/all-mpnet-base-v2
- 4.23 Mihaiii/gte-micro-v4
- 4.18 Mihaiii/Wartortle
- 4.14 Snowflake/snowflake-arctic-embed-l
- 4.13 Mihaiii/Bulbasaur
- 4.03 Snowflake/snowflake-arctic-embed-s
- 3.97 Snowflake/snowflake-arctic-embed-m
- 3.91 Mihaiii/gte-micro
- 3.88 brahmairesearch/slx-v0.1
- 3.88 sentence-transformers/all-MiniLM-L6-v2
- 3.58 minishlab/M2V_base_glove
- 3.57 Mihaiii/Squirtle
- 3.38 Mihaiii/Venusaur
- 3.15 Omartificial-Intelligence-Space/Marbert-all-nli-triplet-Matryoshka
- 3.11 Snowflake/snowflake-arctic-embed-m-long
- 3.10 minishlab/M2V_base_glove_subword
- 3.03 minishlab/potion-base-8M
- 3.01 cointegrated/rubert-tiny
- 2.97 Snowflake/snowflake-arctic-embed-xs
- 2.81 Snowflake/snowflake-arctic-embed-m-v1.5
- 2.47 minishlab/potion-base-4M
- 2.41 consciousAI/cai-lunaris-text-embeddings
- 2.36 minishlab/M2V_base_output
- 2.00 Omartificial-Intelligence-Space/Arabic-mpnet-base-all-nli-triplet
- 1.83 ai-forever/sbert_large_mt_nlu_ru
- 1.66 ai-forever/sbert_large_nlu_ru
- 1.60 minishlab/potion-base-2M
- 1.13 malenia1/ternary-weight-embedding
- 0.98 jinaai/jina-embeddings-v2-small-en
- 0.47 Hum-Works/lodestone-base-4096-v1
- 0.18 Jaume/gemma-2b-embeddings
- 0.12 jinaai/jina-embeddings-v2-base-en
- 1
Gemini Embedding 2 99.5%
- 2 microsoft/harrier-oss-v1-27b 87.7%
- 3 tencent/KaLM-Embedding-Gemma3-12B-2511 85.6%
- 4 intfloat/multilingual-e5-large-instruct 83.7%
- 5 codefuse-ai/F2LLM-v2-14B 82.8%
- 6 facebook/SONAR (Sentence-Level Multimodal and Language-Agnostic Representations) 82.1%
- 7 google/gemini-embedding-001 82%
- 8 nvidia/llama-embed-nemotron-8b 81.5%
- 9 sentence-transformers/LaBSE 81.1%
- 10 codefuse-ai/F2LLM-v2-8B 81%
- 11 OrdalieTech/Solon-embeddings-large-0.1 80.5%
- 12
Qwen3-Embedding-8B 79.7%
- 13 microsoft/harrier-oss-v1-0.6b 79.5%
- 14 Octen/Octen-Embedding-8B 78.9%
- 15 microsoft/harrier-oss-v1-270m 78.4%
- 16 codefuse-ai/F2LLM-v2-4B 77.7%
- 17 Omartificial-Intelligence-Space/Arabic-labse-Matryoshka 76.3%
- 18
Qwen3-Embedding-4B 76.3%
- 19 Lajavaness/bilingual-embedding-large 76.1%
- 20 Bytedance/Seed1.6-embedding-1215 76%
- 21 GritLM/GritLM-8x7B 75.2%
- 22 hotchpotch/bekko-embedding-v1-a25m 74.1%
- 23 BAAI/bge-m3 73.8%
- 24 BidirLM/BidirLM-1B-Embedding 73.8%
- 25 codefuse-ai/F2LLM-v2-1.7B 73.7%
- 26 Alibaba-NLP/gte-Qwen2-7B-instruct 73.5%
- 27 BidirLM/BidirLM-Omni-2.5B-Embedding 72.8%
- 28 Cohere/Cohere-embed-multilingual-v3.0 72.5%
- 29 manu/bge-m3-custom-fr 72.3%
- 30 ai-sage/Giga-Embeddings-instruct-10B-A1.8B-0826 72.3%
- 31 GritLM/GritLM-7B 72.1%
- 32 Mira190/Euler-Legal-Embedding-V1 71.3%
- 33 Linq-AI-Research/Linq-Embed-Mistral 71.3%
- 34 jinaai/jina-embeddings-v3 71.1%
- 35 BidirLM/BidirLM-1.7B-Embedding 71.1%
- 36 Salesforce/SFR-Embedding-Mistral 70.5%
- 37 google/text-multilingual-embedding-002 70.2%
- 38 intfloat/e5-mistral-7b-instruct 70.1%
- 39 Lajavaness/bilingual-embedding-base 69.3%
- 40 intfloat/multilingual-e5-small 68.9%
- 41 intfloat/multilingual-e5-base 68.1%
- 42 ICT-TIME-and-Querit/BOOM_4B_v1 68%
- 43 ICT-TIME-and-Querit/ICT-TIME-and-Querit-embedding-v1 68%
- 44 mongodb/voyage-multilingual-2 67.8%
- 45 hotchpotch/bekko-embedding-v1-a8m 67.6%
- 46 Lajavaness/bilingual-embedding-small 67.2%
- 47 ai-sage/Giga-Embeddings-instruct-3B-0826 67.2%
- 48 Snowflake/snowflake-arctic-embed-l-v2.0 66.8%
- 49 Salesforce/SFR-Embedding-2_R 66%
- 50 openai/text-embedding-3-large 65.6%
- 51 codefuse-ai/F2LLM-v2-0.6B 65.6%
- 52 omarelshehy/arabic-english-sts-matryoshka 64.3%
- 53 Omartificial-Intelligence-Space/Arabic-all-nli-triplet-Matryoshka 63.6%
- 54 nvidia/NV-Embed-v2 63%
- 55 codefuse-ai/F2LLM-v2-330M 62.4%
- 56 mongodb/voyage-3 62.2%
- 57 sentence-transformers/paraphrase-multilingual-mpnet-base-v2 62.2%
- 58 mongodb/voyage-3-lite 61.9%
- 59 jinaai/jina-embeddings-v5-omni-small 61.1%
- 60 jinaai/jina-embeddings-v5-text-small 61.1%
- 61 BidirLM/BidirLM-0.6B-Embedding 60.4%
- 62 Gameselo/STS-multilingual-mpnet-base-v2 59.5%
- 63 mongodb/voyage-3.5 59.3%
- 64 Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet 58.6%
- 65
Qwen3-Embedding-0.6B 58%
- 66 ibm-granite/granite-embedding-278m-multilingual 56.9%
- 67 Alibaba-NLP/gte-Qwen1.5-7B-instruct 56.8%
- 68 sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 56.6%
- 69 jinaai/jina-embeddings-v5-omni-nano 56.5%
- 70 jinaai/jina-embeddings-v5-text-nano 56.5%
- 71 ibm-granite/granite-embedding-311m-multilingual-r2 55.7%
- 72 nomic-ai/nomic-embed-text-v2-moe 54.3%
- 73
EmbeddingGemma 2 54.2%
- 74 Alibaba-NLP/gte-Qwen2-1.5B-instruct 54%
- 75 shibing624/text2vec-base-multilingual 53.7%
- 76 BidirLM/BidirLM-270M-Embedding 53.4%
- 77 Cohere/Cohere-embed-multilingual-light-v3.0 52.7%
- 78 codefuse-ai/F2LLM-v2-160M 51.8%
- 79 google/embeddinggemma-300m 51.4%
- 80 HIT-TMG/KaLM-embedding-multilingual-mini-v1 50.6%
- 81 sentence-transformers/static-similarity-mrl-multilingual-v1 50.2%
- 82 HIT-TMG/KaLM-embedding-multilingual-mini-instruct-v1 50.1%
- 83 ibm-granite/granite-embedding-107m-multilingual 49.3%
- 84 openai/text-embedding-3-small 48.9%
- 85 NovaSearch/jasper_en_vision_language_v1 47.5%
- 86 zeroentropy/zembed-1 46.5%
- 87 codefuse-ai/F2LLM-v2-80M 46.2%
- 88 nvidia/NV-Embed-v1 45.7%
- 89 NovaSearch/stella_en_1.5B_v5 44.2%
- 90 mongodb/voyage-large-2-instruct 42.8%
- 91 Snowflake/snowflake-arctic-embed-m-v2.0 40.8%
- 92 hotchpotch/bekko-static-embedding-v1 39.8%
- 93 ibm-granite/granite-embedding-97m-multilingual-r2 38.8%
- 94 amazon/Titan-text-embeddings-v2 38.3%
- 95 izhx/udever-bloom-7b1 35.2%
- 96 minishlab/potion-multilingual-128M 32.6%
- 97 izhx/udever-bloom-3b 32.5%
- 98 bigscience/sgpt-bloom-7b1-msmarco 29.6%
- 99 izhx/udever-bloom-1b1 28.6%
- 100 cnmoro/static-nomic-384-pten-v2-st 27.6%
- 101 deepfile/embedder-100p 23.1%
- 102 izhx/udever-bloom-560m 23.1%
- 103 sentence-transformers/sentence-t5-xl 21%
- 104 NovaSearch/stella_en_400M_v5 20%
- 105 sentence-transformers/gtr-t5-xl 18.5%
- 106 manu/sentence_croissant_alpha_v0.2 18%
- 107 VPLabs/SearchMap_Preview 17.8%
- 108 manu/sentence_croissant_alpha_v0.3 17.4%
- 109 manu/sentence_croissant_alpha_v0.4 16.7%
- 110 sentence-transformers/sentence-t5-large 16.3%
- 111 intfloat/e5-large-v2 11.7%
- 112 Cohere/Cohere-embed-english-v3.0 11.3%
- 113 Haon-Chen/speed-embedding-7b-instruct 11%
- 114 intfloat/e5-large 9.79%
- 115 intfloat/e5-base-v2 9.06%
- 116 sdadas/mmlw-e5-large 8.76%
- 117 aari1995/German_Semantic_STS_V2 8.17%
- 118 intfloat/e5-base 8.11%
- 119 WhereIsAI/UAE-Large-V1 8.04%
- 120 ibm-granite/granite-embedding-125m-english 7.93%
- 121 sdadas/mmlw-roberta-large 7.89%
- 122 intfloat/e5-small 7.74%
- 123 mixedbread-ai/mxbai-embed-large-v1 7.59%
- 124 ai-forever/ru-en-RoSBERTa 7.57%
- 125 BAAI/bge-large-en-v1.5 7.51%
- 126 nomic-ai/nomic-embed-text-v1-unsupervised 7.49%
- 127 intfloat/e5-small-v2 7.31%
- 128 avsolatorio/GIST-large-Embedding-v0 7.21%
- 129 jinaai/jina-embedding-b-en-v1 7.18%
- 130 sdadas/mmlw-roberta-base 7.12%
- 131 sdadas/mmlw-e5-small 7.09%
- 132 thenlper/gte-large 7.02%
- 133 nomic-ai/nomic-embed-text-v1-ablated 6.89%
- 134 sdadas/mmlw-e5-base 6.87%
- 135 infgrad/stella-base-en-v2 6.71%
- 136 dwzhu/e5-base-4k 6.57%
- 137 avsolatorio/GIST-Embedding-v0 6.39%
- 138 thenlper/gte-base 6.38%
- 139 BAAI/bge-base-en-v1.5 6.34%
- 140 ibm-granite/granite-embedding-30m-english 6.14%
- 141 silma-ai/silma-embeddding-matryoshka-v0.1 6.10%
- 142 Cohere/Cohere-embed-english-light-v3.0 5.89%
- 143 nomic-ai/nomic-embed-text-v1.5 5.83%
- 144 nomic-ai/nomic-embed-text-v1 5.83%
- 145 jinaai/jina-embedding-s-en-v1 5.74%
- 146 BAAI/bge-small-en-v1.5 5.49%
- 147 avsolatorio/NoInstruct-small-Embedding-v0 5.45%
- 148 cointegrated/rubert-tiny2 5.45%
- 149 thenlper/gte-small 5.37%
- 150 avsolatorio/GIST-small-Embedding-v0 5.31%
- 151 abhinand/MedEmbed-small-v0.1 5.21%
- 152 sentence-transformers/all-MiniLM-L12-v2 5.10%
- 153 avsolatorio/GIST-all-MiniLM-L6-v2 4.66%
- 154 Mihaiii/Ivysaur 4.64%
- 155 Omartificial-Intelligence-Space/Arabert-all-nli-triplet-Matryoshka 4.55%
- 156 sentence-transformers/all-mpnet-base-v2 4.40%
- 157 Mihaiii/gte-micro-v4 4.23%
- 158 Mihaiii/Wartortle 4.18%
- 159 Snowflake/snowflake-arctic-embed-l 4.14%
- 160 Mihaiii/Bulbasaur 4.13%
- 161 Snowflake/snowflake-arctic-embed-s 4.03%
- 162 Snowflake/snowflake-arctic-embed-m 3.97%
- 163 Mihaiii/gte-micro 3.91%
- 164 brahmairesearch/slx-v0.1 3.88%
- 165 sentence-transformers/all-MiniLM-L6-v2 3.88%
- 166 minishlab/M2V_base_glove 3.58%
- 167 Mihaiii/Squirtle 3.57%
- 168 Mihaiii/Venusaur 3.38%
- 169 Omartificial-Intelligence-Space/Marbert-all-nli-triplet-Matryoshka 3.15%
- 170 Snowflake/snowflake-arctic-embed-m-long 3.11%
- 171 minishlab/M2V_base_glove_subword 3.10%
- 172 minishlab/potion-base-8M 3.03%
- 173 cointegrated/rubert-tiny 3.01%
- 174 Snowflake/snowflake-arctic-embed-xs 2.97%
- 175 Snowflake/snowflake-arctic-embed-m-v1.5 2.81%
- 176 minishlab/potion-base-4M 2.47%
- 177 consciousAI/cai-lunaris-text-embeddings 2.41%
- 178 minishlab/M2V_base_output 2.36%
- 179 Omartificial-Intelligence-Space/Arabic-mpnet-base-all-nli-triplet 2.00%
- 180 ai-forever/sbert_large_mt_nlu_ru 1.83%
- 181 ai-forever/sbert_large_nlu_ru 1.66%
- 182 minishlab/potion-base-2M 1.60%
- 183 malenia1/ternary-weight-embedding 1.13%
- 184 jinaai/jina-embeddings-v2-small-en 0.98%
- 185 Hum-Works/lodestone-base-4096-v1 0.47%
- 186 Jaume/gemma-2b-embeddings 0.18%
- 187 jinaai/jina-embeddings-v2-base-en 0.12%
Source: MTEB’s leaderboard
- 1
Gemini Embedding 2 99.5%
- 2
Qwen3-Embedding-8B 79.7%
- 3
Qwen3-Embedding-4B 76.3%
- 4
Qwen3-Embedding-0.6B 58%
- 5
EmbeddingGemma 2 54.2%
Source: MTEB’s leaderboard
| # | Entry | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | | 99.5% | Mar 2026 | ai.google.dev | Official leaderboard | |
| 2 | microsoft/harrier-oss-v1-27b open weights, 27B | 87.7% | huggingface.co | Official leaderboard | ||
| 3 | tencent/KaLM-Embedding-Gemma3-12B-2511 open weights, 11.8B | 85.6% | huggingface.co | Official leaderboard | ||
| 4 | intfloat/multilingual-e5-large-instruct open weights, 0.6B | 83.7% | huggingface.co | Official leaderboard | ||
| 5 | codefuse-ai/F2LLM-v2-14B open weights, 14B | 82.8% | huggingface.co | Official leaderboard | ||
| 6 | facebook/SONAR (Sentence-Level Multimodal and Language-Agnostic Representations) open weights | 82.1% | ai.meta.com | Official leaderboard | ||
| 7 | google/gemini-embedding-001 closed weights | 82% | ai.google.dev | Official leaderboard | ||
| 8 | nvidia/llama-embed-nemotron-8b open weights, 7.5B | 81.5% | huggingface.co | Official leaderboard | ||
| 9 | sentence-transformers/LaBSE open weights, 0.5B | 81.1% | huggingface.co | Official leaderboard | ||
| 10 | codefuse-ai/F2LLM-v2-8B open weights, 7.6B | 81% | huggingface.co | Official leaderboard | ||
| 11 | OrdalieTech/Solon-embeddings-large-0.1 open weights, 0.6B | 80.5% | huggingface.co | Official leaderboard | ||
| 12 | | 79.7% | Jun 2025 | huggingface.co | Official leaderboard | |
| 13 | microsoft/harrier-oss-v1-0.6b open weights, 0.6B | 79.5% | huggingface.co | Official leaderboard | ||
| 14 | Octen/Octen-Embedding-8B open weights, 7.6B | 78.9% | huggingface.co | Official leaderboard | ||
| 15 | microsoft/harrier-oss-v1-270m open weights, 0.3B | 78.4% | huggingface.co | Official leaderboard | ||
| 16 | codefuse-ai/F2LLM-v2-4B open weights, 4B | 77.7% | huggingface.co | Official leaderboard | ||
| 17 | Omartificial-Intelligence-Space/Arabic-labse-Matryoshka open weights, 0.5B | 76.3% | huggingface.co | Official leaderboard | ||
| 18 | | 76.3% | Jun 2025 | huggingface.co | Official leaderboard | |
| 19 | Lajavaness/bilingual-embedding-large open weights, 0.6B | 76.1% | huggingface.co | Official leaderboard | ||
| 20 | Bytedance/Seed1.6-embedding-1215 closed weights | 76% | console.volcengine.com | Official leaderboard | ||
| 21 | GritLM/GritLM-8x7B open weights, 57.9B | 75.2% | huggingface.co | Official leaderboard | ||
| 22 | hotchpotch/bekko-embedding-v1-a25m open weights, 0.1B | 74.1% | huggingface.co | Official leaderboard | ||
| 23 | BAAI/bge-m3 open weights, 0.6B | 73.8% | huggingface.co | Official leaderboard | ||
| 24 | BidirLM/BidirLM-1B-Embedding open weights, 1B | 73.8% | huggingface.co | Official leaderboard | ||
| 25 | codefuse-ai/F2LLM-v2-1.7B open weights, 1.7B | 73.7% | huggingface.co | Official leaderboard | ||
| 26 | Alibaba-NLP/gte-Qwen2-7B-instruct open weights, 7.1B | 73.5% | huggingface.co | Official leaderboard | ||
| 27 | BidirLM/BidirLM-Omni-2.5B-Embedding open weights, 2.4B | 72.8% | huggingface.co | Official leaderboard | ||
| 28 | Cohere/Cohere-embed-multilingual-v3.0 closed weights | 72.5% | cohere.com | Official leaderboard | ||
| 29 | manu/bge-m3-custom-fr open weights, 0.6B | 72.3% | huggingface.co | Official leaderboard | ||
| 30 | ai-sage/Giga-Embeddings-instruct-10B-A1.8B-0826 open weights, 10.5B | 72.3% | huggingface.co | Official leaderboard | ||
| 31 | GritLM/GritLM-7B open weights, 7.2B | 72.1% | huggingface.co | Official leaderboard | ||
| 32 | Mira190/Euler-Legal-Embedding-V1 open weights, 7.6B | 71.3% | huggingface.co | Official leaderboard | ||
| 33 | Linq-AI-Research/Linq-Embed-Mistral open weights, 7.1B | 71.3% | huggingface.co | Official leaderboard | ||
| 34 | jinaai/jina-embeddings-v3 open weights, 0.6B | 71.1% | huggingface.co | Official leaderboard | ||
| 35 | BidirLM/BidirLM-1.7B-Embedding open weights, 1.7B | 71.1% | huggingface.co | Official leaderboard | ||
| 36 | Salesforce/SFR-Embedding-Mistral open weights, 7.1B | 70.5% | huggingface.co | Official leaderboard | ||
| 37 | google/text-multilingual-embedding-002 closed weights | 70.2% | cloud.google.com | Official leaderboard | ||
| 38 | intfloat/e5-mistral-7b-instruct open weights, 7.1B | 70.1% | huggingface.co | Official leaderboard | ||
| 39 | Lajavaness/bilingual-embedding-base open weights, 0.3B | 69.3% | huggingface.co | Official leaderboard | ||
| 40 | intfloat/multilingual-e5-small open weights, 0.1B | 68.9% | huggingface.co | Official leaderboard | ||
| 41 | intfloat/multilingual-e5-base open weights, 0.3B | 68.1% | huggingface.co | Official leaderboard | ||
| 42 | ICT-TIME-and-Querit/BOOM_4B_v1 open weights, 4B | 68% | huggingface.co | Official leaderboard | ||
| 43 | ICT-TIME-and-Querit/ICT-TIME-and-Querit-embedding-v1 open weights, 4B | 68% | huggingface.co | Official leaderboard | ||
| 44 | mongodb/voyage-multilingual-2 closed weights | 67.8% | blog.voyageai.com | Official leaderboard | ||
| 45 | hotchpotch/bekko-embedding-v1-a8m open weights, 0.1B | 67.6% | huggingface.co | Official leaderboard | ||
| 46 | Lajavaness/bilingual-embedding-small open weights, 0.1B | 67.2% | huggingface.co | Official leaderboard | ||
| 47 | ai-sage/Giga-Embeddings-instruct-3B-0826 open weights, 3.2B | 67.2% | huggingface.co | Official leaderboard | ||
| 48 | Snowflake/snowflake-arctic-embed-l-v2.0 open weights, 0.6B | 66.8% | huggingface.co | Official leaderboard | ||
| 49 | Salesforce/SFR-Embedding-2_R open weights, 7.1B | 66% | huggingface.co | Official leaderboard | ||
| 50 | openai/text-embedding-3-large closed weights | 65.6% | openai.com | Official leaderboard | ||
| 51 | codefuse-ai/F2LLM-v2-0.6B open weights, 0.6B | 65.6% | huggingface.co | Official leaderboard | ||
| 52 | omarelshehy/arabic-english-sts-matryoshka open weights, 0.6B | 64.3% | huggingface.co | Official leaderboard | ||
| 53 | Omartificial-Intelligence-Space/Arabic-all-nli-triplet-Matryoshka open weights, 0.3B | 63.6% | huggingface.co | Official leaderboard | ||
| 54 | nvidia/NV-Embed-v2 open weights, 7.9B | 63% | huggingface.co | Official leaderboard | ||
| 55 | codefuse-ai/F2LLM-v2-330M open weights, 0.3B | 62.4% | huggingface.co | Official leaderboard | ||
| 56 | mongodb/voyage-3 closed weights | 62.2% | blog.voyageai.com | Official leaderboard | ||
| 57 | sentence-transformers/paraphrase-multilingual-mpnet-base-v2 open weights, 0.3B | 62.2% | huggingface.co | Official leaderboard | ||
| 58 | mongodb/voyage-3-lite closed weights | 61.9% | blog.voyageai.com | Official leaderboard | ||
| 59 | jinaai/jina-embeddings-v5-omni-small open weights, 1.6B | 61.1% | huggingface.co | Official leaderboard | ||
| 60 | jinaai/jina-embeddings-v5-text-small open weights, 0.6B | 61.1% | huggingface.co | Official leaderboard | ||
| 61 | BidirLM/BidirLM-0.6B-Embedding open weights, 0.6B | 60.4% | huggingface.co | Official leaderboard | ||
| 62 | Gameselo/STS-multilingual-mpnet-base-v2 open weights, 0.3B | 59.5% | huggingface.co | Official leaderboard | ||
| 63 | mongodb/voyage-3.5 closed weights | 59.3% | blog.voyageai.com | Official leaderboard | ||
| 64 | Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet open weights, 0.1B | 58.6% | huggingface.co | Official leaderboard | ||
| 65 | | 58% | Jun 2025 | huggingface.co | Official leaderboard | |
| 66 | ibm-granite/granite-embedding-278m-multilingual open weights, 0.3B | 56.9% | huggingface.co | Official leaderboard | ||
| 67 | Alibaba-NLP/gte-Qwen1.5-7B-instruct open weights, 7.1B | 56.8% | huggingface.co | Official leaderboard | ||
| 68 | sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 open weights, 0.1B | 56.6% | huggingface.co | Official leaderboard | ||
| 69 | jinaai/jina-embeddings-v5-omni-nano open weights, 1B | 56.5% | huggingface.co | Official leaderboard | ||
| 70 | jinaai/jina-embeddings-v5-text-nano open weights, 0.2B | 56.5% | huggingface.co | Official leaderboard | ||
| 71 | ibm-granite/granite-embedding-311m-multilingual-r2 open weights, 0.3B | 55.7% | huggingface.co | Official leaderboard | ||
| 72 | nomic-ai/nomic-embed-text-v2-moe open weights, 0.5B | 54.3% | huggingface.co | Official leaderboard | ||
| 73 | | 54.2% | Oct 2026 | huggingface.co | Official leaderboard | |
| 74 | Alibaba-NLP/gte-Qwen2-1.5B-instruct open weights, 1.5B | 54% | huggingface.co | Official leaderboard | ||
| 75 | shibing624/text2vec-base-multilingual open weights, 0.1B | 53.7% | huggingface.co | Official leaderboard | ||
| 76 | BidirLM/BidirLM-270M-Embedding open weights, 0.3B | 53.4% | huggingface.co | Official leaderboard | ||
| 77 | Cohere/Cohere-embed-multilingual-light-v3.0 closed weights | 52.7% | cohere.com | Official leaderboard | ||
| 78 | codefuse-ai/F2LLM-v2-160M open weights, 0.2B | 51.8% | huggingface.co | Official leaderboard | ||
| 79 | google/embeddinggemma-300m open weights, 0.3B | 51.4% | ai.google.dev | Official leaderboard | ||
| 80 | HIT-TMG/KaLM-embedding-multilingual-mini-v1 open weights, 0.5B | 50.6% | huggingface.co | Official leaderboard | ||
| 81 | sentence-transformers/static-similarity-mrl-multilingual-v1 open weights, 0.1B | 50.2% | huggingface.co | Official leaderboard | ||
| 82 | HIT-TMG/KaLM-embedding-multilingual-mini-instruct-v1 open weights, 0.5B | 50.1% | huggingface.co | Official leaderboard | ||
| 83 | ibm-granite/granite-embedding-107m-multilingual open weights, 0.1B | 49.3% | huggingface.co | Official leaderboard | ||
| 84 | openai/text-embedding-3-small closed weights | 48.9% | openai.com | Official leaderboard | ||
| 85 | NovaSearch/jasper_en_vision_language_v1 open weights, 1.6B | 47.5% | huggingface.co | Official leaderboard | ||
| 86 | zeroentropy/zembed-1 open weights, 4B | 46.5% | huggingface.co | Official leaderboard | ||
| 87 | codefuse-ai/F2LLM-v2-80M open weights, 0.1B | 46.2% | huggingface.co | Official leaderboard | ||
| 88 | nvidia/NV-Embed-v1 open weights, 7.9B | 45.7% | huggingface.co | Official leaderboard | ||
| 89 | NovaSearch/stella_en_1.5B_v5 open weights, 1.5B | 44.2% | huggingface.co | Official leaderboard | ||
| 90 | mongodb/voyage-large-2-instruct closed weights | 42.8% | blog.voyageai.com | Official leaderboard | ||
| 91 | Snowflake/snowflake-arctic-embed-m-v2.0 open weights, 0.3B | 40.8% | huggingface.co | Official leaderboard | ||
| 92 | hotchpotch/bekko-static-embedding-v1 open weights, 0.2B | 39.8% | huggingface.co | Official leaderboard | ||
| 93 | ibm-granite/granite-embedding-97m-multilingual-r2 open weights, 0.1B | 38.8% | huggingface.co | Official leaderboard | ||
| 94 | amazon/Titan-text-embeddings-v2 closed weights | 38.3% | huggingface.co | Official leaderboard | ||
| 95 | izhx/udever-bloom-7b1 open weights | 35.2% | huggingface.co | Official leaderboard | ||
| 96 | minishlab/potion-multilingual-128M open weights, 0.1B | 32.6% | huggingface.co | Official leaderboard | ||
| 97 | izhx/udever-bloom-3b open weights, 3B | 32.5% | huggingface.co | Official leaderboard | ||
| 98 | bigscience/sgpt-bloom-7b1-msmarco open weights, 7.1B | 29.6% | huggingface.co | Official leaderboard | ||
| 99 | izhx/udever-bloom-1b1 open weights, 1.1B | 28.6% | huggingface.co | Official leaderboard | ||
| 100 | cnmoro/static-nomic-384-pten-v2-st open weights, 0.1B | 27.6% | huggingface.co | Official leaderboard | ||
| 101 | deepfile/embedder-100p open weights, 0.3B | 23.1% | huggingface.co | Official leaderboard | ||
| 102 | izhx/udever-bloom-560m open weights, 0.6B | 23.1% | huggingface.co | Official leaderboard | ||
| 103 | sentence-transformers/sentence-t5-xl open weights, 1.2B | 21% | huggingface.co | Official leaderboard | ||
| 104 | NovaSearch/stella_en_400M_v5 open weights, 0.4B | 20% | huggingface.co | Official leaderboard | ||
| 105 | sentence-transformers/gtr-t5-xl open weights, 1.2B | 18.5% | huggingface.co | Official leaderboard | ||
| 106 | manu/sentence_croissant_alpha_v0.2 open weights, 1.3B | 18% | huggingface.co | Official leaderboard | ||
| 107 | VPLabs/SearchMap_Preview open weights, 0.4B | 17.8% | huggingface.co | Official leaderboard | ||
| 108 | manu/sentence_croissant_alpha_v0.3 open weights, 1.3B | 17.4% | huggingface.co | Official leaderboard | ||
| 109 | manu/sentence_croissant_alpha_v0.4 open weights, 1.3B | 16.7% | huggingface.co | Official leaderboard | ||
| 110 | sentence-transformers/sentence-t5-large open weights, 0.3B | 16.3% | huggingface.co | Official leaderboard | ||
| 111 | intfloat/e5-large-v2 open weights, 0.3B | 11.7% | huggingface.co | Official leaderboard | ||
| 112 | Cohere/Cohere-embed-english-v3.0 closed weights | 11.3% | cohere.com | Official leaderboard | ||
| 113 | Haon-Chen/speed-embedding-7b-instruct open weights, 7.1B | 11% | huggingface.co | Official leaderboard | ||
| 114 | intfloat/e5-large open weights, 0.3B | 9.79% | huggingface.co | Official leaderboard | ||
| 115 | intfloat/e5-base-v2 open weights, 0.1B | 9.06% | huggingface.co | Official leaderboard | ||
| 116 | sdadas/mmlw-e5-large open weights, 0.6B | 8.76% | huggingface.co | Official leaderboard | ||
| 117 | aari1995/German_Semantic_STS_V2 open weights, 0.3B | 8.17% | huggingface.co | Official leaderboard | ||
| 118 | intfloat/e5-base open weights, 0.1B | 8.11% | huggingface.co | Official leaderboard | ||
| 119 | WhereIsAI/UAE-Large-V1 open weights, 0.3B | 8.04% | huggingface.co | Official leaderboard | ||
| 120 | ibm-granite/granite-embedding-125m-english open weights, 0.1B | 7.93% | huggingface.co | Official leaderboard | ||
| 121 | sdadas/mmlw-roberta-large open weights, 0.4B | 7.89% | huggingface.co | Official leaderboard | ||
| 122 | intfloat/e5-small open weights, 0B | 7.74% | huggingface.co | Official leaderboard | ||
| 123 | mixedbread-ai/mxbai-embed-large-v1 open weights, 0.3B | 7.59% | huggingface.co | Official leaderboard | ||
| 124 | ai-forever/ru-en-RoSBERTa open weights, 0.4B | 7.57% | huggingface.co | Official leaderboard | ||
| 125 | BAAI/bge-large-en-v1.5 open weights, 0.3B | 7.51% | huggingface.co | Official leaderboard | ||
| 126 | nomic-ai/nomic-embed-text-v1-unsupervised open weights, 0.1B | 7.49% | huggingface.co | Official leaderboard | ||
| 127 | intfloat/e5-small-v2 open weights, 0B | 7.31% | huggingface.co | Official leaderboard | ||
| 128 | avsolatorio/GIST-large-Embedding-v0 open weights, 0.3B | 7.21% | huggingface.co | Official leaderboard | ||
| 129 | jinaai/jina-embedding-b-en-v1 open weights, 0.1B | 7.18% | huggingface.co | Official leaderboard | ||
| 130 | sdadas/mmlw-roberta-base open weights, 0.1B | 7.12% | huggingface.co | Official leaderboard | ||
| 131 | sdadas/mmlw-e5-small open weights, 0.1B | 7.09% | huggingface.co | Official leaderboard | ||
| 132 | thenlper/gte-large open weights, 0.3B | 7.02% | huggingface.co | Official leaderboard | ||
| 133 | nomic-ai/nomic-embed-text-v1-ablated open weights, 0.1B | 6.89% | huggingface.co | Official leaderboard | ||
| 134 | sdadas/mmlw-e5-base open weights, 0.3B | 6.87% | huggingface.co | Official leaderboard | ||
| 135 | infgrad/stella-base-en-v2 open weights, 0.1B | 6.71% | huggingface.co | Official leaderboard | ||
| 136 | dwzhu/e5-base-4k open weights, 0.1B | 6.57% | huggingface.co | Official leaderboard | ||
| 137 | avsolatorio/GIST-Embedding-v0 open weights, 0.1B | 6.39% | huggingface.co | Official leaderboard | ||
| 138 | thenlper/gte-base open weights, 0.1B | 6.38% | huggingface.co | Official leaderboard | ||
| 139 | BAAI/bge-base-en-v1.5 open weights, 0.1B | 6.34% | huggingface.co | Official leaderboard | ||
| 140 | ibm-granite/granite-embedding-30m-english open weights, 0B | 6.14% | huggingface.co | Official leaderboard | ||
| 141 | silma-ai/silma-embeddding-matryoshka-v0.1 open weights, 0.1B | 6.10% | huggingface.co | Official leaderboard | ||
| 142 | Cohere/Cohere-embed-english-light-v3.0 closed weights | 5.89% | cohere.com | Official leaderboard | ||
| 143 | nomic-ai/nomic-embed-text-v1.5 open weights, 0.1B | 5.83% | huggingface.co | Official leaderboard | ||
| 144 | nomic-ai/nomic-embed-text-v1 open weights, 0.1B | 5.83% | huggingface.co | Official leaderboard | ||
| 145 | jinaai/jina-embedding-s-en-v1 open weights, 0B | 5.74% | huggingface.co | Official leaderboard | ||
| 146 | BAAI/bge-small-en-v1.5 open weights, 0B | 5.49% | huggingface.co | Official leaderboard | ||
| 147 | avsolatorio/NoInstruct-small-Embedding-v0 open weights, 0B | 5.45% | huggingface.co | Official leaderboard | ||
| 148 | cointegrated/rubert-tiny2 open weights, 0B | 5.45% | huggingface.co | Official leaderboard | ||
| 149 | thenlper/gte-small open weights, 0B | 5.37% | huggingface.co | Official leaderboard | ||
| 150 | avsolatorio/GIST-small-Embedding-v0 open weights, 0B | 5.31% | huggingface.co | Official leaderboard | ||
| 151 | abhinand/MedEmbed-small-v0.1 open weights, 0B | 5.21% | huggingface.co | Official leaderboard | ||
| 152 | sentence-transformers/all-MiniLM-L12-v2 open weights, 0B | 5.10% | huggingface.co | Official leaderboard | ||
| 153 | avsolatorio/GIST-all-MiniLM-L6-v2 open weights, 0B | 4.66% | huggingface.co | Official leaderboard | ||
| 154 | Mihaiii/Ivysaur open weights, 0B | 4.64% | huggingface.co | Official leaderboard | ||
| 155 | Omartificial-Intelligence-Space/Arabert-all-nli-triplet-Matryoshka open weights, 0.1B | 4.55% | huggingface.co | Official leaderboard | ||
| 156 | sentence-transformers/all-mpnet-base-v2 open weights, 0.1B | 4.40% | huggingface.co | Official leaderboard | ||
| 157 | Mihaiii/gte-micro-v4 open weights, 0B | 4.23% | huggingface.co | Official leaderboard | ||
| 158 | Mihaiii/Wartortle open weights, 0B | 4.18% | huggingface.co | Official leaderboard | ||
| 159 | Snowflake/snowflake-arctic-embed-l open weights, 0.3B | 4.14% | huggingface.co | Official leaderboard | ||
| 160 | Mihaiii/Bulbasaur open weights, 0B | 4.13% | huggingface.co | Official leaderboard | ||
| 161 | Snowflake/snowflake-arctic-embed-s open weights, 0B | 4.03% | huggingface.co | Official leaderboard | ||
| 162 | Snowflake/snowflake-arctic-embed-m open weights, 0.1B | 3.97% | huggingface.co | Official leaderboard | ||
| 163 | Mihaiii/gte-micro open weights, 0B | 3.91% | huggingface.co | Official leaderboard | ||
| 164 | brahmairesearch/slx-v0.1 open weights, 0B | 3.88% | huggingface.co | Official leaderboard | ||
| 165 | sentence-transformers/all-MiniLM-L6-v2 open weights, 0B | 3.88% | huggingface.co | Official leaderboard | ||
| 166 | minishlab/M2V_base_glove open weights, 0.1B | 3.58% | huggingface.co | Official leaderboard | ||
| 167 | Mihaiii/Squirtle open weights, 0B | 3.57% | huggingface.co | Official leaderboard | ||
| 168 | Mihaiii/Venusaur open weights, 0B | 3.38% | huggingface.co | Official leaderboard | ||
| 169 | Omartificial-Intelligence-Space/Marbert-all-nli-triplet-Matryoshka open weights, 0.2B | 3.15% | huggingface.co | Official leaderboard | ||
| 170 | Snowflake/snowflake-arctic-embed-m-long open weights, 0.1B | 3.11% | huggingface.co | Official leaderboard | ||
| 171 | minishlab/M2V_base_glove_subword open weights, 0.1B | 3.10% | huggingface.co | Official leaderboard | ||
| 172 | minishlab/potion-base-8M open weights, 0B | 3.03% | huggingface.co | Official leaderboard | ||
| 173 | cointegrated/rubert-tiny open weights, 0B | 3.01% | huggingface.co | Official leaderboard | ||
| 174 | Snowflake/snowflake-arctic-embed-xs open weights, 0B | 2.97% | huggingface.co | Official leaderboard | ||
| 175 | Snowflake/snowflake-arctic-embed-m-v1.5 open weights, 0.1B | 2.81% | huggingface.co | Official leaderboard | ||
| 176 | minishlab/potion-base-4M open weights, 0B | 2.47% | huggingface.co | Official leaderboard | ||
| 177 | consciousAI/cai-lunaris-text-embeddings open weights, 0.3B | 2.41% | huggingface.co | Official leaderboard | ||
| 178 | minishlab/M2V_base_output open weights, 0B | 2.36% | huggingface.co | Official leaderboard | ||
| 179 | Omartificial-Intelligence-Space/Arabic-mpnet-base-all-nli-triplet open weights, 0.1B | 2.00% | huggingface.co | Official leaderboard | ||
| 180 | ai-forever/sbert_large_mt_nlu_ru open weights, 0.4B | 1.83% | huggingface.co | Official leaderboard | ||
| 181 | ai-forever/sbert_large_nlu_ru open weights, 0.4B | 1.66% | huggingface.co | Official leaderboard | ||
| 182 | minishlab/potion-base-2M open weights, 0B | 1.60% | huggingface.co | Official leaderboard | ||
| 183 | malenia1/ternary-weight-embedding open weights, 0.1B | 1.13% | huggingface.co | Official leaderboard | ||
| 184 | jinaai/jina-embeddings-v2-small-en open weights, 0B | 0.98% | huggingface.co | Official leaderboard | ||
| 185 | Hum-Works/lodestone-base-4096-v1 open weights, 0.1B | 0.47% | huggingface.co | Official leaderboard | ||
| 186 | Jaume/gemma-2b-embeddings open weights, 2.5B | 0.18% | huggingface.co | Official leaderboard | ||
| 187 | jinaai/jina-embeddings-v2-base-en open weights, 0.1B | 0.12% | huggingface.co | Official leaderboard |
| # | Model | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | | 99.5% | Mar 2026 | ai.google.dev | Official leaderboard | |
| 2 | | 79.7% | Jun 2025 | huggingface.co | Official leaderboard | |
| 3 | | 76.3% | Jun 2025 | huggingface.co | Official leaderboard | |
| 4 | | 58% | Jun 2025 | huggingface.co | Official leaderboard | |
| 5 | | 54.2% | Oct 2026 | huggingface.co | Official leaderboard |
Rankings by part
We read 112 parts from MTEB’s leaderboard, each ranking models on one part of the test, such as afr-eng, amh-eng and ang-eng. Gemini Embedding 2 leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.
afr-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.7%
- 3
Qwen3-Embedding-4B 93.2%
- 4
Qwen3-Embedding-0.6B 65.3%
- 5
EmbeddingGemma 2 63.2%
amh-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 69.5%
- 3
Qwen3-Embedding-4B 54.5%
- 4
EmbeddingGemma 2 ↑1 38.1%
- 5
Qwen3-Embedding-0.6B ↓1 33%
ang-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 90.2%
- 3
Qwen3-Embedding-4B 83.1%
- 4
Qwen3-Embedding-0.6B 62.6%
- 5
EmbeddingGemma 2 35.6%
ara-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 92.3%
- 3
Qwen3-Embedding-4B 91.6%
- 4
Qwen3-Embedding-0.6B 80.2%
- 5
EmbeddingGemma 2 79.1%
arq-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 63.6%
- 3
Qwen3-Embedding-4B 55.5%
- 4
Qwen3-Embedding-0.6B 29.9%
- 5
EmbeddingGemma 2 28.4%
arz-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 80.8%
- 3
Qwen3-Embedding-4B 76.3%
- 4
EmbeddingGemma 2 ↑1 64.3%
- 5
Qwen3-Embedding-0.6B ↓1 54.4%
ast-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 83.3%
- 3
Qwen3-Embedding-4B 81.3%
- 4
Qwen3-Embedding-0.6B 73.6%
- 5
EmbeddingGemma 2 71.6%
awa-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 83.3%
- 3
Qwen3-Embedding-4B 76%
- 4
Qwen3-Embedding-0.6B 54.7%
- 5
EmbeddingGemma 2 44.9%
aze-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 89.8%
- 3
Qwen3-Embedding-4B 88.7%
- 4
Qwen3-Embedding-0.6B 61%
- 5
EmbeddingGemma 2 59.4%
bel-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.3%
- 3
Qwen3-Embedding-4B 91.1%
- 4
Qwen3-Embedding-0.6B 66.6%
- 5
EmbeddingGemma 2 64.7%
ben-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 88%
- 3
Qwen3-Embedding-4B 86.1%
- 4
EmbeddingGemma 2 ↑1 75.6%
- 5
Qwen3-Embedding-0.6B ↓1 68.6%
ber-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 95.8%
- 2
Qwen3-Embedding-8B 11.2%
- 3
Qwen3-Embedding-4B 9.09%
- 4
Qwen3-Embedding-0.6B 7.53%
- 5
EmbeddingGemma 2 6.04%
bos-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.9%
- 3
Qwen3-Embedding-4B 91%
- 4
EmbeddingGemma 2 ↑1 85%
- 5
Qwen3-Embedding-0.6B ↓1 75%
bre-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 16%
- 3
Qwen3-Embedding-4B 14.1%
- 4
Qwen3-Embedding-0.6B 9.70%
- 5
EmbeddingGemma 2 7.83%
bul-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.7%
- 3
Qwen3-Embedding-4B 93.4%
- 4
Qwen3-Embedding-0.6B 86.4%
- 5
EmbeddingGemma 2 80.5%
cat-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.5%
- 3
Qwen3-Embedding-4B 92.8%
- 4
Qwen3-Embedding-0.6B 79.6%
- 5
EmbeddingGemma 2 74.7%
cbk-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 82.7%
- 3
Qwen3-Embedding-8B ↓1 82.6%
- 4
EmbeddingGemma 2 ↑1 73.1%
- 5
Qwen3-Embedding-0.6B ↓1 66.7%
ceb-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 56.1%
- 3
Qwen3-Embedding-4B 46%
- 4
EmbeddingGemma 2 ↑1 18.5%
- 5
Qwen3-Embedding-0.6B ↓1 17.3%
ces-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 96.1%
- 3
Qwen3-Embedding-4B 96%
- 4
EmbeddingGemma 2 ↑1 89.5%
- 5
Qwen3-Embedding-0.6B ↓1 82.4%
cha-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 48.1%
- 3
Qwen3-Embedding-4B 47.6%
- 4
Qwen3-Embedding-0.6B 43.2%
- 5
EmbeddingGemma 2 36.6%
cmn-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 96.4%
- 3
Qwen3-Embedding-4B 96.2%
- 4
Qwen3-Embedding-0.6B 95.3%
- 5
EmbeddingGemma 2 93.6%
cor-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 9.56%
- 3
Qwen3-Embedding-4B 8.52%
- 4
Qwen3-Embedding-0.6B 7.26%
- 5
EmbeddingGemma 2 5.39%
csb-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 68.8%
- 3
Qwen3-Embedding-4B 67.4%
- 4
Qwen3-Embedding-0.6B 46.1%
- 5
EmbeddingGemma 2 31.9%
cym-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 76.2%
- 3
Qwen3-Embedding-4B 60.6%
- 4
Qwen3-Embedding-0.6B 17.4%
- 5
EmbeddingGemma 2 11.1%
dan-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 95.5%
- 3
Qwen3-Embedding-4B 94.9%
- 4
Qwen3-Embedding-0.6B 87%
- 5
EmbeddingGemma 2 81.9%
deu-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 99.7%
- 3
Qwen3-Embedding-4B 99.5%
- 4
Qwen3-Embedding-0.6B 98.5%
- 5
EmbeddingGemma 2 98%
dsb-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 67.2%
- 3
Qwen3-Embedding-4B 66.2%
- 4
Qwen3-Embedding-0.6B 43.1%
- 5
EmbeddingGemma 2 38.5%
dtp-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 87.9%
- 2
Qwen3-Embedding-8B 12.4%
- 3
Qwen3-Embedding-4B 11.5%
- 4
Qwen3-Embedding-0.6B 9.57%
- 5
EmbeddingGemma 2 7.09%
ell-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.9%
- 3
Qwen3-Embedding-4B 90.9%
- 4
Qwen3-Embedding-0.6B 78.2%
- 5
EmbeddingGemma 2 72.4%
epo-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.4%
- 3
Qwen3-Embedding-4B 93.8%
- 4
Qwen3-Embedding-0.6B 71.8%
- 5
EmbeddingGemma 2 41.1%
est-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 78.7%
- 3
Qwen3-Embedding-8B ↓1 78.5%
- 4
Qwen3-Embedding-0.6B 34.2%
- 5
EmbeddingGemma 2 29.5%
eus-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 67.7%
- 3
Qwen3-Embedding-4B 55.2%
- 4
Qwen3-Embedding-0.6B 21.1%
- 5
EmbeddingGemma 2 18%
fao-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 86.8%
- 3
Qwen3-Embedding-4B 83%
- 4
Qwen3-Embedding-0.6B 48%
- 5
EmbeddingGemma 2 32%
fin-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 90.8%
- 3
Qwen3-Embedding-8B ↓1 90.2%
- 4
EmbeddingGemma 2 ↑1 65.5%
- 5
Qwen3-Embedding-0.6B ↓1 62.7%
fra-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 95.5%
- 3
Qwen3-Embedding-8B ↓1 95.3%
- 4
Qwen3-Embedding-0.6B 94.2%
- 5
EmbeddingGemma 2 91.7%
fry-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 83.6%
- 3
Qwen3-Embedding-4B 78.9%
- 4
Qwen3-Embedding-0.6B 59.3%
- 5
EmbeddingGemma 2 45.4%
gla-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 97.4%
- 2
Qwen3-Embedding-8B 58.9%
- 3
Qwen3-Embedding-4B 38.9%
- 4
Qwen3-Embedding-0.6B 11.7%
- 5
EmbeddingGemma 2 4.04%
gle-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 76.7%
- 3
Qwen3-Embedding-4B 53.3%
- 4
Qwen3-Embedding-0.6B 13.1%
- 5
EmbeddingGemma 2 5.64%
glg-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.2%
- 3
Qwen3-Embedding-4B 91.7%
- 4
EmbeddingGemma 2 ↑1 81.3%
- 5
Qwen3-Embedding-0.6B ↓1 80.9%
gsw-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 59.7%
- 3
Qwen3-Embedding-4B 55.1%
- 4
Qwen3-Embedding-0.6B 44.9%
- 5
EmbeddingGemma 2 42.2%
heb-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 90%
- 3
Qwen3-Embedding-8B ↓1 89.9%
- 4
Qwen3-Embedding-0.6B 71.5%
- 5
EmbeddingGemma 2 61.5%
hin-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 97.5%
- 3
Qwen3-Embedding-8B ↓1 97.5%
- 4
EmbeddingGemma 2 ↑1 93.3%
- 5
Qwen3-Embedding-0.6B ↓1 88.9%
hrv-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 94.7%
- 3
Qwen3-Embedding-8B ↓1 94.3%
- 4
Qwen3-Embedding-0.6B 80.2%
- 5
EmbeddingGemma 2 79.8%
hsb-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 78.5%
- 3
Qwen3-Embedding-4B 77.2%
- 4
Qwen3-Embedding-0.6B 50.9%
- 5
EmbeddingGemma 2 44.6%
hun-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.1%
- 3
Qwen3-Embedding-4B 91.1%
- 4
Qwen3-Embedding-0.6B 72.2%
- 5
EmbeddingGemma 2 64.6%
hye-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.3%
- 3
Qwen3-Embedding-4B 92%
- 4
Qwen3-Embedding-0.6B 60.8%
- 5
EmbeddingGemma 2 36%
ido-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 85%
- 3
Qwen3-Embedding-4B 84.2%
- 4
Qwen3-Embedding-0.6B 68.1%
- 5
EmbeddingGemma 2 53%
ile-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 88.8%
- 3
Qwen3-Embedding-8B ↓1 86.9%
- 4
Qwen3-Embedding-0.6B 75%
- 5
EmbeddingGemma 2 65.3%
ina-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 95%
- 3
Qwen3-Embedding-8B ↓1 94.8%
- 4
Qwen3-Embedding-0.6B 90.5%
- 5
EmbeddingGemma 2 83.6%
ind-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.4%
- 3
Qwen3-Embedding-4B 93.4%
- 4
EmbeddingGemma 2 ↑1 89.3%
- 5
Qwen3-Embedding-0.6B ↓1 89.1%
isl-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.7%
- 3
Qwen3-Embedding-4B 86.8%
- 4
Qwen3-Embedding-0.6B 43.2%
- 5
EmbeddingGemma 2 23.1%
ita-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.2%
- 3
Qwen3-Embedding-4B 93%
- 4
Qwen3-Embedding-0.6B 88.6%
- 5
EmbeddingGemma 2 88.3%
jav-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 78.8%
- 3
Qwen3-Embedding-4B 63.7%
- 4
Qwen3-Embedding-0.6B 31.7%
- 5
EmbeddingGemma 2 27.6%
jpn-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 96.2%
- 3
Qwen3-Embedding-8B ↓1 96.1%
- 4
Qwen3-Embedding-0.6B 91.6%
- 5
EmbeddingGemma 2 91.3%
kab-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 94.3%
- 2
Qwen3-Embedding-8B 6.13%
- 3
Qwen3-Embedding-4B 5.16%
- 4
Qwen3-Embedding-0.6B 3.74%
- 5
EmbeddingGemma 2 1.67%
kat-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 88%
- 3
Qwen3-Embedding-4B 82.6%
- 4
EmbeddingGemma 2 ↑1 59.9%
- 5
Qwen3-Embedding-0.6B ↓1 48.8%
kaz-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 84.9%
- 3
Qwen3-Embedding-4B 81.9%
- 4
Qwen3-Embedding-0.6B 48%
- 5
EmbeddingGemma 2 37.5%
khm-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 75%
- 3
Qwen3-Embedding-4B 67.5%
- 4
EmbeddingGemma 2 ↑1 55.4%
- 5
Qwen3-Embedding-0.6B ↓1 29.7%
kor-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 91.9%
- 3
Qwen3-Embedding-4B 91.3%
- 4
EmbeddingGemma 2 ↑1 86.5%
- 5
Qwen3-Embedding-0.6B ↓1 83.8%
kur-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 53.3%
- 3
Qwen3-Embedding-4B 44.7%
- 4
Qwen3-Embedding-0.6B 25.8%
- 5
EmbeddingGemma 2 17.4%
kzj-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 65%
- 2
Qwen3-Embedding-8B 15%
- 3
Qwen3-Embedding-4B 14.3%
- 4
Qwen3-Embedding-0.6B 11.7%
- 5
EmbeddingGemma 2 7.85%
lat-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 82.2%
- 3
Qwen3-Embedding-4B 78%
- 4
Qwen3-Embedding-0.6B 52.2%
- 5
EmbeddingGemma 2 30.3%
lfn-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 74.9%
- 3
Qwen3-Embedding-8B ↓1 74.8%
- 4
Qwen3-Embedding-0.6B 61.5%
- 5
EmbeddingGemma 2 58%
lit-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 90%
- 3
Qwen3-Embedding-4B 89.3%
- 4
Qwen3-Embedding-0.6B 58.8%
- 5
EmbeddingGemma 2 49.7%
lvs-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 92%
- 3
Qwen3-Embedding-4B 90.5%
- 4
Qwen3-Embedding-0.6B 57.1%
- 5
EmbeddingGemma 2 36%
mal-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 98.3%
- 3
Qwen3-Embedding-4B 97.3%
- 4
EmbeddingGemma 2 ↑1 86.4%
- 5
Qwen3-Embedding-0.6B ↓1 77.3%
mar-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 90.8%
- 3
Qwen3-Embedding-4B 89.1%
- 4
EmbeddingGemma 2 ↑1 76%
- 5
Qwen3-Embedding-0.6B ↓1 66.6%
max-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 69.6%
- 3
Qwen3-Embedding-8B ↓1 67.4%
- 4
Qwen3-Embedding-0.6B 48.7%
- 5
EmbeddingGemma 2 46.9%
mhr-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 17.8%
- 3
Qwen3-Embedding-8B ↓1 17.1%
- 4
Qwen3-Embedding-0.6B 14.4%
- 5
EmbeddingGemma 2 10.1%
mkd-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 89.5%
- 3
Qwen3-Embedding-8B ↓1 89.4%
- 4
Qwen3-Embedding-0.6B 66.3%
- 5
EmbeddingGemma 2 53.8%
mon-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 84.4%
- 3
Qwen3-Embedding-4B 77.4%
- 4
Qwen3-Embedding-0.6B 33.2%
- 5
EmbeddingGemma 2 27.4%
nds-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 84.8%
- 3
Qwen3-Embedding-4B 79.6%
- 4
Qwen3-Embedding-0.6B 57%
- 5
EmbeddingGemma 2 49.2%
nld-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 97.4%
- 3
Qwen3-Embedding-4B 97%
- 4
EmbeddingGemma 2 ↑1 93.6%
- 5
Qwen3-Embedding-0.6B ↓1 91.6%
nno-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 95.9%
- 3
Qwen3-Embedding-4B 93.9%
- 4
Qwen3-Embedding-0.6B 78.6%
- 5
EmbeddingGemma 2 66.1%
nob-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 98.9%
- 3
Qwen3-Embedding-4B 98.5%
- 4
Qwen3-Embedding-0.6B 90%
- 5
EmbeddingGemma 2 88%
nov-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 81.4%
- 3
Qwen3-Embedding-8B ↓1 80.8%
- 4
Qwen3-Embedding-0.6B 65.2%
- 5
EmbeddingGemma 2 63%
oci-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 68.4%
- 3
Qwen3-Embedding-4B 65.9%
- 4
Qwen3-Embedding-0.6B 45.1%
- 5
EmbeddingGemma 2 40.9%
orv-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 67.2%
- 3
Qwen3-Embedding-4B 65.3%
- 4
Qwen3-Embedding-0.6B 37.1%
- 5
EmbeddingGemma 2 24.2%
pam-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 17.5%
- 3
Qwen3-Embedding-4B 15.5%
- 4
Qwen3-Embedding-0.6B 10.9%
- 5
EmbeddingGemma 2 8.57%
pes-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 94%
- 3
Qwen3-Embedding-8B ↓1 93.7%
- 4
EmbeddingGemma 2 ↑1 85.6%
- 5
Qwen3-Embedding-0.6B ↓1 82.3%
pms-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 74.7%
- 3
Qwen3-Embedding-4B 73.6%
- 4
Qwen3-Embedding-0.6B 51.2%
- 5
EmbeddingGemma 2 37.7%
pol-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 97.3%
- 3
Qwen3-Embedding-4B 96.6%
- 4
EmbeddingGemma 2 ↑1 93.5%
- 5
Qwen3-Embedding-0.6B ↓1 91.7%
por-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.9%
- 3
Qwen3-Embedding-4B 94.6%
- 4
EmbeddingGemma 2 ↑1 93.5%
- 5
Qwen3-Embedding-0.6B ↓1 93.2%
ron-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.7%
- 3
Qwen3-Embedding-4B 93.5%
- 4
Qwen3-Embedding-0.6B 84.7%
- 5
EmbeddingGemma 2 79.7%
rus-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.3%
- 3
Qwen3-Embedding-4B 93.6%
- 4
Qwen3-Embedding-0.6B 92.4%
- 5
EmbeddingGemma 2 91.8%
slk-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93%
- 3
Qwen3-Embedding-4B 92.3%
- 4
EmbeddingGemma 2 ↑1 81.4%
- 5
Qwen3-Embedding-0.6B ↓1 76.4%
slv-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 88.6%
- 3
Qwen3-Embedding-4B 85.9%
- 4
EmbeddingGemma 2 ↑1 67.7%
- 5
Qwen3-Embedding-0.6B ↓1 67.7%
spa-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 98.7%
- 3
Qwen3-Embedding-4B 98.5%
- 4
Qwen3-Embedding-0.6B 96.2%
- 5
EmbeddingGemma 2 96.1%
sqi-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 88.7%
- 3
Qwen3-Embedding-4B 83.1%
- 4
Qwen3-Embedding-0.6B 49.5%
- 5
EmbeddingGemma 2 29.3%
srp-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.4%
- 3
Qwen3-Embedding-4B 93.4%
- 4
Qwen3-Embedding-0.6B 74.2%
- 5
EmbeddingGemma 2 67.9%
swe-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 95.7%
- 3
Qwen3-Embedding-4B 95.4%
- 4
Qwen3-Embedding-0.6B 84.4%
- 5
EmbeddingGemma 2 82.6%
swg-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 84.6%
- 3
Qwen3-Embedding-4B 79.4%
- 4
Qwen3-Embedding-0.6B 54.6%
- 5
EmbeddingGemma 2 43.8%
swh-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 62.4%
- 3
Qwen3-Embedding-4B 49.6%
- 4
EmbeddingGemma 2 ↑1 30.4%
- 5
Qwen3-Embedding-0.6B ↓1 27.5%
tam-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 93.6%
- 3
Qwen3-Embedding-4B 91.3%
- 4
EmbeddingGemma 2 ↑1 77.6%
- 5
Qwen3-Embedding-0.6B ↓1 66%
tat-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 73.5%
- 3
Qwen3-Embedding-4B 61.4%
- 4
Qwen3-Embedding-0.6B 25.2%
- 5
EmbeddingGemma 2 16.8%
tel-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.9%
- 3
Qwen3-Embedding-4B 94.4%
- 4
EmbeddingGemma 2 ↑1 77.7%
- 5
Qwen3-Embedding-0.6B ↓1 71.6%
tgl-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 91.2%
- 3
Qwen3-Embedding-4B 85.8%
- 4
EmbeddingGemma 2 ↑1 47.4%
- 5
Qwen3-Embedding-0.6B ↓1 38.5%
tha-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 96.8%
- 3
Qwen3-Embedding-4B 96%
- 4
EmbeddingGemma 2 ↑1 93%
- 5
Qwen3-Embedding-0.6B ↓1 91.6%
tuk-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 53.2%
- 3
Qwen3-Embedding-4B 42.1%
- 4
Qwen3-Embedding-0.6B 25.8%
- 5
EmbeddingGemma 2 17.4%
tur-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 96.8%
- 3
Qwen3-Embedding-4B 94.9%
- 4
EmbeddingGemma 2 ↑1 90%
- 5
Qwen3-Embedding-0.6B ↓1 80.6%
tzl-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 60.7%
- 3
Qwen3-Embedding-8B ↓1 54.6%
- 4
Qwen3-Embedding-0.6B 47.7%
- 5
EmbeddingGemma 2 33.4%
uig-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 78.4%
- 3
Qwen3-Embedding-4B 64.4%
- 4
Qwen3-Embedding-0.6B 12.7%
- 5
EmbeddingGemma 2 4.33%
ukr-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 94.3%
- 3
Qwen3-Embedding-4B 93.5%
- 4
EmbeddingGemma 2 ↑1 89%
- 5
Qwen3-Embedding-0.6B ↓1 87.3%
urd-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 92.4%
- 3
Qwen3-Embedding-4B 91.8%
- 4
Qwen3-Embedding-0.6B 73%
- 5
EmbeddingGemma 2 72.4%
uzb-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 76.6%
- 3
Qwen3-Embedding-4B 68.3%
- 4
Qwen3-Embedding-0.6B 34.9%
- 5
EmbeddingGemma 2 25.6%
vie-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 97.6%
- 3
Qwen3-Embedding-4B 97.4%
- 4
Qwen3-Embedding-0.6B 95.6%
- 5
EmbeddingGemma 2 95.3%
war-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 55.5%
- 3
Qwen3-Embedding-4B 46%
- 4
Qwen3-Embedding-0.6B 21%
- 5
EmbeddingGemma 2 16.6%
wuu-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 94.3%
- 3
Qwen3-Embedding-8B ↓1 94.2%
- 4
Qwen3-Embedding-0.6B 88.4%
- 5
EmbeddingGemma 2 83.5%
xho-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 40.5%
- 3
Qwen3-Embedding-4B 24.9%
- 4
Qwen3-Embedding-0.6B 20.1%
- 5
EmbeddingGemma 2 11.6%
yid-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 85.4%
- 3
Qwen3-Embedding-4B 59%
- 4
Qwen3-Embedding-0.6B 10.9%
- 5
EmbeddingGemma 2 5.45%
yue-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-4B ↑1 93%
- 3
Qwen3-Embedding-8B ↓1 92.7%
- 4
Qwen3-Embedding-0.6B 86.1%
- 5
EmbeddingGemma 2 79.4%
zsm-eng
5 models we track have a score here.
- 1
Gemini Embedding 2 100%
- 2
Qwen3-Embedding-8B 96.1%
- 3
Qwen3-Embedding-4B 95.9%
- 4
Qwen3-Embedding-0.6B 91.2%
- 5
EmbeddingGemma 2 89%
2 models have held number one on MTEB’s leaderboard since 2023
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: MTEB’s leaderboard
Gemini Embedding 2 on the official leaderboard passed 85% in March 2026
That was 87 months after its release. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: MTEB’s leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Multilingual
- MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
- WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET.
K-EXAONE 90.5%
- MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language.
Qwen3.7-Max 87%
- CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories.
Solar Open 2 90.7%
- Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
- MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%
Data audit
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (187 entries).
- Pass Models scored 5 models we track, plus 182 other entries on its leaderboard.
- Pass Newest score The newest is from October 2026, for EmbeddingGemma 2.
- Note Frontier models None of its models are on the capabilities index.
- Note Headroom Saturated. The best is 99.5%, past 85% since March 2026.
Leaderboard import 3 pass
- Pass Importer We read the board on MTEB: 5 rows for models we track and 182 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.