SEA-HELM Leaderboard

English Performance

Note

Results for all models, except AISG models, are sourced from the MTEB website and have not been re-validated.

English Scores by Model

Qwen3 Embedding

68.71

EmbeddingGemma

0.3B

65.11

Qwen3 Embedding

0.6B

64.72

E5 Mistral Instruct (e5-mistral-7b-instruct)

GritLM

63.22

jina-embeddings-v3

0.6B

62.36

Cohere embed-multilingual-v3.0

61.43

SEA-LION Embedding E5

0.6B

61.41

Multilingual E5 Instruct (multilingual-e5-large-instruct)

0.6B

61.21

SEA-LION ModernBERT Embedding

0.6B

60.64

Multilingual E5 (multilingual-e5-large)

0.6B

58.66

SEA-LION ModernBERT Embedding

0.3B

58.21

voyage-3

57.25

paraphrase-multilingual-mpnet-base-v2

0.3B

55.71

paraphrase-multilingual-MiniLM-L12-v2

0.1B

53.99

LaBSE

0.5B

49.45

BGE-Multilingual-Gemma2

BGE-M3

text-embedding-3-small

English Competencies

Model	EN	Pair Classification	Classification	Semantic Textual Similarity	Retrieval	Clustering	Reranking	Summarization
Qwen3 Embedding 8B Alibaba	68.71	87.52	90.43	88.58	69.44	58.57	51.56	34.83
EmbeddingGemma 0.3B Google	65.11	87.29	87.55	83.61	55.69	56.55	47.43	37.64
Qwen3 Embedding 0.6B Alibaba	64.72	84.37	84.58	86.57	61.83	54.05	48.18	33.43
E5 Mistral Instruct (e5-mistral-7b-instruct) 7B Microsoft	64.00	88.42	79.85	84.32	57.62	51.44	49.78	36.57
GritLM 7B GritLM	63.22	87.29	81.25	83.03	54.95	50.82	49.59	35.65
jina-embeddings-v3 0.6B Jina AI	62.36	84.02	84.12	85.82	54.29	47.41	47.94	32.90
Cohere embed-multilingual-v3.0 CohereLabs	61.43	86.15	79.26	83.05	55.69	48.05	47.36	30.43
SEA-LION Embedding E5 0.6B AISG	61.41	87.35	77.82	84.76	50.24	51.05	47.08	31.59
Multilingual E5 Instruct (multilingual-e5-large-instruct) 0.6B Microsoft	61.21	86.24	75.54	84.72	53.47	49.89	48.74	29.89
SEA-LION ModernBERT Embedding 0.6B AISG	60.64	85.37	79.00	83.84	50.06	50.32	45.57	30.29
Multilingual E5 (multilingual-e5-large) 0.6B Microsoft	58.66	84.75	74.74	81.39	51.53	42.07	44.74	31.41
SEA-LION ModernBERT Embedding 0.3B AISG	58.21	83.24	77.64	81.79	44.80	48.45	44.57	26.98
voyage-3 Voyage AI	57.25	78.55	69.72	78.59	53.46	44.85	46.72	28.87
paraphrase-multilingual-mpnet-base-v2 0.3B UKP Lab	55.71	81.73	72.50	80.50	37.14	43.52	45.16	29.45
paraphrase-multilingual-MiniLM-L12-v2 0.1B UKP Lab	53.99	80.01	69.19	78.73	35.93	41.74	45.43	26.92
LaBSE 0.5B UKP Lab	49.45	78.87	67.98	70.63	18.05	36.08	41.28	33.27
BGE-Multilingual-Gemma2 9B Beijing Academy of Artificial Intelligence	—	85.97	88.74	85.39	57.15	—	48.19	—
BGE-M3 6B Beijing Academy of Artificial Intelligence	—	91.61	73.02	80.61	51.04	40.96	—	34.77
text-embedding-3-small OpenAI	—	—	77.56	81.35	53.48	47.46	—	32.49

No records