← Catalog
lightonai/Reason-ModernColBERT Add to compare Open comparison →
Primitive: /score · Score ·
ModernBERT
Reason-ModernColBERT is a late interaction model trained on the reasonir-hq dataset. It achieves extremely competitive performance on the BRIGHT benchmark aimed at evaluating reasoning-intensive retrieval performance, outperforming all existing models up to 7B (more than 45 times its size) and ev...
Long context
View on Hugging Face → Fine-tuned from lightonai/GTE-ModernColBERT-v1
Overview
Hardware:
L4 RTX-PRO-6000 — drives latency, throughput & cost
Size 149M params Tasks /encode · /score License cc-by-nc-4.0 Languages en Latency — Throughput — Cost — /1M tok
Cost is approximate — computed from list GPU prices; your actual price depends on the provider you deploy SIE with.
Scoring Inputs text Max sequence length 8,192
Benchmarks Duplicate question detection from AskUbuntu
Corpus: 6,743 Queries: 360
default
Quality
ndcg at 10 0.6338
map at 10 0.4783
mrr at 10 0.7156
muvera
Quality
ndcg at 10 0.6235
map at 10 0.4634
mrr at 10 0.7084
Reference →
Chinese medical question answering reranking (v1)
Corpus: 100,000 Queries: 2,000
default
Quality
ndcg at 10 0.4804
map at 10 0.4118
mrr at 10 0.4980
muvera
Quality
ndcg at 10 0.2331
map at 10 0.1694
mrr at 10 0.2333
Reference →
Chinese medical question answering reranking (v2)
Corpus: 108,000 Queries: 4,000
default
Quality
ndcg at 10 0.5024
map at 10 0.4346
mrr at 10 0.5171
muvera
Quality
ndcg at 10 0.2451
map at 10 0.1851
mrr at 10 0.2454
Reference →
Duplicate question retrieval from StackExchange Physics
Corpus: 38,314 Queries: 1,039
muvera_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.1321
mrr at 10 0.1620
ndcg at 10 0.1681
default_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.3875
mrr at 10 0.4489
ndcg at 10 0.4442
Reference →
Code search with natural language queries
Corpus: 6,267 Queries: 500
muvera_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.1637
mrr at 10 0.2028
ndcg at 10 0.2187
default_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.3020
mrr at 10 0.2829
ndcg at 10 0.3896
Reference →
Financial opinion mining and question answering
Corpus: 57,599 Queries: 648
muvera_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.0486
mrr at 10 0.0953
ndcg at 10 0.0701
default_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.3880
mrr at 10 0.5504
ndcg at 10 0.4672
Reference →
Question answering on consumer contracts
Corpus: 153 Queries: 396
muvera_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.4025
mrr at 10 0.4024
ndcg at 10 0.4732
default_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.7882
mrr at 10 0.7895
ndcg at 10 0.8328
Reference →
Multilingual MARCO passage reranking (Chinese)
default
Quality
ndcg at 10 0.2002
map at 10 0.1512
mrr at 10 0.1536
muvera
Quality
ndcg at 10 0.0714
map at 10 0.0554
mrr at 10 0.0570
Reference →
Biomedical literature search from NutritionFacts.org
Corpus: 3,593 Queries: 323
muvera_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.0596
mrr at 10 0.3258
ndcg at 10 0.1853
default_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.1342
mrr at 10 0.5719
ndcg at 10 0.3582
Reference →
Citation prediction, document classification, and recommendation for scientific papers
Corpus: 25,656 Queries: 1,000
muvera_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.0370
mrr at 10 0.1317
ndcg at 10 0.0733
default_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.1200
mrr at 10 0.3406
ndcg at 10 0.2015
Reference →
Scientific claim verification using research literature
Corpus: 5,183 Queries: 300
muvera_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.3902
mrr at 10 0.4020
ndcg at 10 0.4383
default_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.7107
mrr at 10 0.7226
ndcg at 10 0.7531
Reference →
Programming question answering from Stack Overflow
Corpus: 19,931 Queries: 1,994
muvera_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.1952
mrr at 10 0.1952
ndcg at 10 0.2346
default_candidates-k-50_candidates-model-Alibaba-NLP__gte-multilingual-base
Quality
map at 10 0.6559
mrr at 10 0.6559
ndcg at 10 0.6893
Reference →
