The blended average the public MTEB leaderboard shows across a model's task-type scores; the number most people mean when they say 'MTEB score'.
unassessed
| Category | embedding |
|---|---|
| Subcategory | text embedding (blended average) |
| Page status | active |
| Metric | mean score (Task Mean or Task Type Mean) |
| Direction | higher_is_better |
| Unit | % |
| Dataset licence | Varies by underlying dataset; the MTEB benchmark code is Apache-2.0. |
| Publisher | Hugging Face and Cohere (original authors); maintained today by the open embeddings-benchmark community |
mteb_overall is not a task of its own. It is the summary column the public MTEB leaderboard shows by default, combining every task type a given leaderboard variant covers — retrieval, classification, clustering, reranking, STS, summarization, pair classification, and on some variants bitext mining — into one number per model. It is what most vendor announcements and comparison charts quote when they say "MTEB."
Not a task: an aggregation (mean) of a model's scores on every task in a given MTEB benchmark variant, computed after the model has already been scored on each individual task.
| Model | Provider | Score | Card as of |
|---|---|---|---|
| bge reranker v2 m3 | BAAI | 70.1 | 2026-04 |
| NV Embed v2 | NVIDIA | 69.5 | 2026-04 |
| jina reranker v2 base multilingual | Jina AI | 68.5 | 2026-04 |
| bge m3 | BAAI | 68.2 | 2026-04 |
| Qwen3 Embedding 8B | Alibaba / Qwen Team | 68.0 | 2026-04 |
| Qwen3 VL Embedding 8B | Alibaba / Qwen Team | 68.0 | 2026-04 |
| jina embeddings v4 | Jina AI | 67.8 | 2026-04 |
| jina embeddings v4 vllm retrieval | Jina AI | 67.8 | 2026-04 |
| Voyage 3 | Voyage AI | 67.2 | 2026-04 |
| bge multilingual gemma2 | BAAI | 66.8 | 2026-04 |
| e5 mistral 7B instruct | intfloat | 66.8 | 2026-04 |
| jina embeddings v3 | Jina AI | 66.5 | 2026-04 |
| Gemini Embedding 001 | Google DeepMind | 66.2 | 2026-04 |
| nomic embed text v2 moe | Nomic AI | 65.5 | 2026-04 |
| nomic embed text v2 moe GGUF | Nomic AI | 65.5 | 2026-04 |
| Qwen3 Embedding 4B | Alibaba / Qwen Team | 65.5 | 2026-04 |
| snowflake arctic embed l v2.0 | Snowflake | 65.5 | 2026-04 |
| SFR Embedding 2 R | Salesforce | 65.2 | 2026-04 |
| text-embedding-3-large | OpenAI | 64.8 | 2026-04 |
| bge large zh v1.5 | BAAI | 64.5 | 2026-04 |
| multilingual e5 large instruct | intfloat | 64.5 | 2026-04 |
| bge large en v1.5 | BAAI | 64.2 | 2026-04 |
| bge large zh | BAAI | 63.5 | 2026-04 |
| granite embedding english r2 | IBM | 63.5 | 2026-04 |
| granite embedding small english r2 | IBM | 63.5 | 2026-04 |
| Voyage Code 3 | Voyage AI | 63.5 | 2026-04 |
| snowflake arctic embed l | Snowflake | 63.2 | 2026-04 |
| bge large en | BAAI | 63.0 | 2026-04 |
| Voyage Finance 2 | Voyage AI | 63.0 | 2026-04 |
| nomic embed text v1.5 | Nomic AI | 62.8 | 2026-04 |
| nomic embed text v1.5 GGUF | Nomic AI | 62.8 | 2026-04 |
| snowflake arctic embed m v2.0 | Snowflake | 62.8 | 2026-04 |
| Voyage Multilingual 2 | Voyage AI | 62.8 | 2026-04 |
| e5 large v2 | intfloat | 62.5 | 2026-04 |
| Voyage Law 2 | Voyage AI | 62.5 | 2026-04 |
| bge base en v1.5 | BAAI | 62.1 | 2026-04 |
| Voyage 3 Lite | Voyage AI | 62.1 | 2026-04 |
| bge base zh v1.5 | BAAI | 62 | 2026-04 |
| granite embedding 278M multilingual | IBM | 61.5 | 2026-04 |
| nomic embed text v1 | Nomic AI | 61.5 | 2026-04 |
| snowflake arctic embed m v1.5 | Snowflake | 61.5 | 2026-04 |
| multilingual e5 large | intfloat | 61.2 | 2026-04 |
| bge base zh | BAAI | 61 | 2026-04 |
| e5 large | intfloat | 61.0 | 2026-04 |
| snowflake arctic embed m long | Snowflake | 61.0 | 2026-04 |
| bge base en | BAAI | 60.8 | 2026-04 |
| jina embeddings v2 base en | Jina AI | 60.8 | 2026-04 |
| bge code v1 | BAAI | 60.5 | 2026-04 |
| nomic embed code | Nomic AI | 60.5 | 2026-04 |
| snowflake arctic embed m | Snowflake | 60.5 | 2026-04 |
| e5 large unsupervised | intfloat | 60.2 | 2026-04 |
| jina embeddings v2 base de | Jina AI | 60.2 | 2026-04 |
| Mistral Embed | Mistral AI | 60.2 | 2026-04 |
| Qwen3 Embedding 0.6B | Alibaba / Qwen Team | 60.0 | 2026-04 |
| Qwen3 VL Embedding 2B | Alibaba / Qwen Team | 60.0 | 2026-04 |
| e5 base v2 | intfloat | 59.8 | 2026-04 |
| jina embeddings v2 base code | Jina AI | 59.5 | 2026-04 |
| SFR Embedding Code 400M R | Salesforce | 59.5 | 2026-04 |
| bge small en v1.5 | BAAI | 59.3 | 2026-04 |
| bge small zh v1.5 | BAAI | 59 | 2026-04 |
| multilingual e5 base | intfloat | 59.0 | 2026-04 |
| e5 base | intfloat | 58.5 | 2026-04 |
| granite embedding 125M english | IBM | 58.5 | 2026-04 |
| jina embeddings v2 small en | Jina AI | 58.5 | 2026-04 |
| text-embedding-3-small | OpenAI | 58.5 | 2026-04 |
| bge small zh | BAAI | 58.2 | 2026-04 |
| granite embedding 107M multilingual | IBM | 58.0 | 2026-04 |
| all mpnet base v2 | Sentence Transformers | 57.8 | 2026-04 |
| snowflake arctic embed s | Snowflake | 57.8 | 2026-04 |
| bge small en | BAAI | 57.5 | 2026-04 |
| e5 small v2 | intfloat | 57.5 | 2026-04 |
| multilingual e5 small | intfloat | 57.2 | 2026-04 |
| all MiniLM L12 v2 | Sentence Transformers | 56.8 | 2026-04 |
| all MiniLM L6 v2 | Sentence Transformers | 56.3 | 2026-04 |
| e5 small | intfloat | 56.2 | 2026-04 |
| all roberta large v1 | Sentence Transformers | 56.0 | 2026-04 |
| multi qa mpnet base dot v1 | Sentence Transformers | 55.8 | 2026-04 |
| text-embedding-ada-002 | OpenAI | 55.8 | 2026-04 |
| multi qa mpnet base cos v1 | Sentence Transformers | 55.5 | 2026-04 |
| paraphrase mpnet base v2 | Sentence Transformers | 55.5 | 2026-04 |
| all distilroberta v1 | Sentence Transformers | 55.2 | 2026-04 |
| snowflake arctic embed xs | Snowflake | 55.2 | 2026-04 |
| paraphrase multilingual mpnet base v2 | Sentence Transformers | 54.8 | 2026-04 |
| multi qa MiniLM L6 cos v1 | Sentence Transformers | 54.2 | 2026-04 |
| paraphrase multilingual MiniLM L12 v2 | Sentence Transformers | 53.5 | 2026-04 |
| paraphrase MiniLM L12 v2 | Sentence Transformers | 53.2 | 2026-04 |
| granite embedding 30M english | IBM | 53.0 | 2026-04 |
| granite embedding 30M sparse | IBM | 53.0 | 2026-04 |
| msmarco bert base dot v5 | Sentence Transformers | 53.0 | 2026-04 |
| paraphrase MiniLM L6 v2 | Sentence Transformers | 52.8 | 2026-04 |
| msmarco MiniLM L12 cos v5 | Sentence Transformers | 52.5 | 2026-04 |
| LaBSE | Sentence Transformers | 52.0 | 2026-04 |
| msmarco MiniLM L6 v3 | Sentence Transformers | 51.5 | 2026-04 |
| paraphrase MiniLM L3 v2 | Sentence Transformers | 50.5 | 2026-04 |
| distiluse base multilingual cased v2 | Sentence Transformers | 50.0 | 2026-04 |
| distiluse base multilingual cased v1 | Sentence Transformers | 48.5 | 2026-04 |