skip to content
gigarouter gigarouter
models / embeddings · coming soon

GTE Large EN v1.5

Alibaba-NLP/gte-large-en-v1.5

published Apr 2024 · updated Aug 2024

GTE Large EN v1.5 is a text embedding model that supports long context up to 8192 tokens and achieves state-of-the-art performance on the MTEB benchmark.

est. price
~$0.008
/ 1M tokens · estimated, set at launch
API providers
0
downloads / mo
1.1M
license
apache-2.0

specs

TaskText Embeddings
ArchitectureTransformer++ (BERT + RoPE + GLU)
Parameters434M
LicenseApache-2.0

about this model

gte-large-en-v1.5 is an English text embedding model that generates dense vector representations of text, supporting a context length of up to 8192 tokens. It is built on a Transformer++ encoder backbone (BERT with Rotary Position Embedding and Gated Linear Units) and is developed by the Institute for Intelligent Computing, Alibaba Group. The model is released under the Apache-2.0 license and its underlying mGTE paper has been accepted to the EMNLP 2024 Industry Track.

Key Strengths and Benchmark Results

The model achieves a state-of-the-art average score of 65.39 on the MTEB benchmark (56 tasks) within its size category, and a competitive 86.71 on the LoCo long-context retrieval benchmark. Its embedding dimension is 1024.

MTEB Benchmark Comparison

Results from the MTEB leaderboard (evaluation using mteb==1.2.0, fp16 auto mix precision, max_length=8192, ntk scaling factor of 2):

Model NameParam Size (M)DimensionSequence LengthAverage (56)Class. (12)Clust. (11)Pair Class. (3)Reran. (4)Retr. (15)STS (10)Summ. (1)
gte-large-en-v1.54091024819265.3977.7547.9584.6358.5057.9181.4330.91
mxbai-embed-large-v1335102451264.6875.6446.7187.260.1154.398532.71
multilingual-e5-large-instruct560102451464.4177.5647.186.1958.5852.4784.7830.39
bge-large-en-v1.5335102451264.2375.9746.0887.1260.0354.2983.1131.61
gte-base-en-v1.5137768819264.1177.1746.8285.3357.6654.0981.9731.17
bge-base-en-v1.510976851263.5575.5345.7786.5558.8653.2582.431.07

LoCo Long-Context Retrieval

Model NameDimensionSequence LengthAverage (5)QsmsumRetrievalSummScreenRetrievalQasperAbstractRetrievalQasperTitleRetrievalGovReportRetrieval
gte-large-v1.51024819286.7144.5592.6199.8297.8198.74
gte-base-v1.5768819287.4449.9191.7899.8297.1398.58
gte-qwen1.5-7b40963276887.5749.3793.1099.6797.5498.21

For further details, see the mGTE paper and GTE paper.

best for

FAQ

What is the maximum sequence length supported by GTE Large EN v1.5?

It supports up to 8192 tokens.

How does this model compare to other large embedding models like BGE Large v1.5?

It achieves 65.39 on MTEB (56 tasks), outperforming BGE Large v1.5 (64.23) and has a longer context length (8192 vs 512).

What is the license for GTE Large EN v1.5?

It is released under the Apache-2.0 license.

How can I call this model via the gigarouter API?

Use the OpenAI-compatible endpoint with your API key, sending text inputs and receiving normalized embeddings.

Does the model support acceleration with xformers?

Yes, it supports unpadding and xformers for faster inference; set trust_remote_code=True and use float16.

not yet live

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.

related embeddings models

compare all →