gte-large-en-v1.5
Alibaba-NLP/gte-large-en-v1.5
A popular open embeddings model, with 1.1M downloads a month. gigarouter benchmarks and hosts it as an OpenAI-compatible API.
about this model
Alibaba-NLP/gte-large-en-v1.5 is an English text embedding model that generates dense vector representations for tasks such as retrieval, clustering, classification, and semantic similarity. It supports a maximum sequence length of 8192 tokens and produces 1024-dimensional embeddings.
Key strengths
- State-of-the-art performance on the MTEB benchmark within the 400M parameter size category.
- Competitive results on the LoCo long-context retrieval benchmark, demonstrating strong handling of extended documents.
- Built on a transformer++ encoder backbone (BERT + RoPE + GLU) optimized for long contexts.
Benchmark results
MTEB (English, 56 tasks): average score 65.39. LoCo (long-context retrieval, 5 tasks): average score 86.71.
| Model | Params (M) | Dim. | Max length | MTEB avg. | LoCo avg. |
|---|---|---|---|---|---|
| gte-large-en-v1.5 | 434 | 1024 | 8192 | 65.39 | 86.71 |
| mxbai-embed-large-v1 | 335 | 1024 | 512 | 64.68 | — |
| bge-large-en-v1.5 | 335 | 1024 | 512 | 64.23 | — |
MTEB subtask breakdown (gte-large-en-v1.5): Classification 77.75, Clustering 47.95, Pair Classification 84.63, Reranking 58.50, Retrieval 57.91, STS 81.43, Summarization 30.91.
Best use cases
- Long-document retrieval and search, where queries and documents exceed 512 tokens.
- High‑accuracy semantic similarity, clustering, and classification tasks in English.
- Applications requiring a lightweight (434M parameter) model with competitive MTEB scores.
We're benchmarking and onboarding gte-large-en-v1.5 as a hosted, OpenAI-compatible API. Sign in for free credit and be ready when it lands, or tell us you want it and we'll prioritize it.