--- title: lightonai/GTE-ModernColBERT-v1 (Score) description: This is a PyLate model trained on the ms-marco-en-bge-gemma dataset. It maps sentences & paragraphs to sequences of 128-dimensional dense ve. ModernBERT, 305M parameters. canonical_url: https://superlinked.com/models/lightonai-gte-moderncolbert-v1--score last_updated: 2026-05-24 --- # lightonai/GTE-ModernColBERT-v1 (Score) This is a PyLate model trained on the ms-marco-en-bge-gemma dataset. It maps sentences & paragraphs to sequences of 128-dimensional dense vectors and can be used for semantic textual similarity using the MaxSim operator. Source: [lightonai/GTE-ModernColBERT-v1 on HuggingFace](https://huggingface.co/lightonai/GTE-ModernColBERT-v1) Base model: [Alibaba-NLP/gte-modernbert-base](https://huggingface.co/Alibaba-NLP/gte-modernbert-base) ## Overview | Field | Value | |-------|-------| | Architecture | ModernBERT | | Parameters | 305M | | Tasks | Encode, Score | | Outputs | Multi-Vec | | Dimensions | Multi-Vec: 128 | | Max sequence length | 8,192 tokens | | License | apache-2.0 | | Inputs | text | ## Benchmarks ### AskUbuntuDupQuestions Domain: technology · Task: reranking · Language: en Duplicate question detection from AskUbuntu Corpus: 6,743 · Queries: 360 **Quality:** ndcg at 10: 0.6388 · map at 10: 0.4817 · mrr at 10: 0.7277 [Reference](https://github.com/taolei87/askubuntu) ### CMedQAv1-reranking Domain: general · Task: reranking · Language: en **Quality:** ndcg at 10: 0.4673 · map at 10: 0.4037 · mrr at 10: 0.4918 ### CMedQAv2-reranking Domain: general · Task: reranking · Language: en **Quality:** ndcg at 10: 0.4864 · map at 10: 0.4193 · mrr at 10: 0.5036 ### CQADupstackPhysicsRetrieval Domain: scientific · Task: retrieval · Language: en Duplicate question retrieval from StackExchange Physics Corpus: 38,314 · Queries: 1,039 **Performance (L4-SPOT b1 c16):** Corpus 1.9K tok/s · Corpus p50 509.4ms · Query 131 tok/s · Query p50 573.4ms **Performance (L4 b1 c16):** Corpus 1.9K tok/s · Corpus p50 509.4ms · Query 131 tok/s · Query p50 573.4ms [Reference](http://nlp.cis.unimelb.edu.au/resources/cqadupstack/) ### CosQA Domain: technology · Task: retrieval · Language: en Code search with natural language queries Corpus: 6,267 · Queries: 500 **Performance (L4-SPOT b1 c16):** Corpus 890 tok/s · Corpus p50 454.2ms · Query 75 tok/s · Query p50 566.6ms **Performance (L4 b1 c16):** Corpus 890 tok/s · Corpus p50 454.2ms · Query 75 tok/s · Query p50 566.6ms [Reference](https://arxiv.org/abs/2105.13239) ### FiQA2018 Domain: finance · Task: retrieval · Language: en Financial opinion mining and question answering Corpus: 57,599 · Queries: 648 **Performance (L4-SPOT b1 c16):** Corpus 2.6K tok/s · Corpus p50 469.6ms · Query 303 tok/s · Query p50 278.2ms **Performance (L4 b1 c16):** Corpus 2.6K tok/s · Corpus p50 469.6ms · Query 303 tok/s · Query p50 278.2ms [Reference](https://sites.google.com/view/fiqa/) ### LegalBenchConsumerContractsQA Domain: legal · Task: retrieval · Language: en Question answering on consumer contracts Corpus: 153 · Queries: 396 **Performance (L4-SPOT b1 c16):** Corpus 6.2K tok/s · Corpus p50 532.8ms · Query 278 tok/s · Query p50 327.3ms **Performance (L4 b1 c16):** Corpus 6.2K tok/s · Corpus p50 532.8ms · Query 278 tok/s · Query p50 327.3ms [Reference](https://huggingface.co/datasets/nguha/legalbench) ### MMarcoReranking Domain: general · Task: reranking · Language: zh Multilingual MARCO passage reranking (Chinese) **Quality:** ndcg at 10: 0.2453 · map at 10: 0.2024 · mrr at 10: 0.2072 [Reference](https://arxiv.org/abs/2304.03679) ### NFCorpus Domain: medical · Task: retrieval · Language: en Biomedical literature search from NutritionFacts.org Corpus: 3,593 · Queries: 323 **Performance (L4-SPOT b1 c16):** Corpus 4.4K tok/s · Corpus p50 463.3ms · Query 111 tok/s · Query p50 299.7ms **Performance (L4 b1 c16):** Corpus 4.4K tok/s · Corpus p50 463.3ms · Query 111 tok/s · Query p50 299.7ms [Reference](https://www.cl.uni-heidelberg.de/statnlpgroup/nfcorpus/) ### SCIDOCS Domain: scientific · Task: retrieval · Language: en Citation prediction, document classification, and recommendation for scientific papers Corpus: 25,656 · Queries: 1,000 **Performance (L4-SPOT b1 c16):** Corpus 4.4K tok/s · Corpus p50 257.6ms · Query 184 tok/s · Query p50 327.2ms **Performance (L4 b1 c16):** Corpus 4.4K tok/s · Corpus p50 257.6ms · Query 184 tok/s · Query p50 327.2ms [Reference](https://allenai.org/data/scidocs) ### SciFact Domain: scientific · Task: retrieval · Language: en Scientific claim verification using research literature Corpus: 5,183 · Queries: 300 **Performance (L4-SPOT b1 c16):** Corpus 9.2K tok/s · Corpus p50 241.6ms · Query 396 tok/s · Query p50 265.9ms **Performance (L4 b1 c16):** Corpus 9.2K tok/s · Corpus p50 241.6ms · Query 396 tok/s · Query p50 265.9ms [Reference](https://github.com/allenai/scifact) ### StackOverflowQA Domain: technology · Task: retrieval · Language: en Programming question answering from Stack Overflow Corpus: 19,931 · Queries: 1,994 **Performance (L4-SPOT b1 c16):** Corpus 3.8K tok/s · Corpus p50 458.1ms · Query 9.2K tok/s · Query p50 222.9ms **Performance (L4 b1 c16):** Corpus 3.8K tok/s · Corpus p50 458.1ms · Query 9.2K tok/s · Query p50 222.9ms [Reference](https://arxiv.org/abs/2407.02883)