Alibaba-NLP/gte-reranker-modernbert-base
Primitive: /score · Score ·
ModernBERT
We are excited to introduce the `gte-modernbert` series of models, which are built upon the latest modernBERT pre-trained encoder-only foundation models. The `gte-modernbert` series models include both text embedding models and rerank models.
Long context
View on Hugging Face → Fine-tuned from answerdotai/ModernBERT-base
Overview
Hardware: — drives latency, throughput & cost
| Size | 150M params |
|---|---|
| Tasks | /score |
| License | apache-2.0 |
| Languages | en |
| Latency | 55 ms |
| Throughput | 11.0K tok/s |
| Cost | $0.020 /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
AskUbuntuDupQuestions
Duplicate question detection from AskUbuntu
Corpus: 6,743 Queries: 360
Quality
ndcg at 10 0.6701
map at 10 0.5148
mrr at 10 0.7570
Performance L4 b1 c16
Query 6.2K tok/s
Query p50 41.9ms
CMedQAv1Reranking
Chinese medical question answering reranking (v1)
Corpus: 100,000 Queries: 2,000
Quality
map at 10 0.4989
mrr at 10 0.5905
CMedQAv2Reranking
Chinese medical question answering reranking (v2)
Corpus: 108,000 Queries: 4,000
Quality
map at 10 0.5024
mrr at 10 0.5880
MMarcoReranking
Multilingual MARCO passage reranking (Chinese)
Quality
map at 10 0.2271
mrr at 10 0.2373
Performance L4 b1 c16
Compare (0)Compare →