GTE-Base
110M BERT-based text embedding model that beat ada-002 on MTEB.
About
GTE-Base is a general text embedding model from Alibaba DAMO Academy, part of the GTE family trained with multi-stage contrastive learning on large-scale relevance text pairs. With 110M parameters and 768 embedding dimensions, it scored around 62.4 on MTEB, outperforming OpenAI's text-embedding-ada-002 despite being roughly ten times smaller.
Released under Apache 2.0 on Hugging Face, it is a standard choice for retrieval, semantic similarity, reranking, and RAG pipelines where a compact, permissively licensed embedder matters. The GTE paper appeared on arXiv in August 2023.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 132; Trending score 72/100 from researcher assessment.