Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
Alibaba-NLPhuggingface.co · Low evidenceAlibaba-NLP's gte-reranker-modernbert-base is part of the gte-modernbert series, offering English text embedding and text reranking models built on modernBERT for retrieval tasks. The model is a 149M parameter encoder/reranker (8192 max input tokens) developed by Tongyi Lab / Alibaba Group and has been evaluated on MTEB, LoCo, and COIR benchmarks. It can be used with Hugging Face transformers or sentence-transformers, supports optional Flash Attention 2, and can be deployed via Text Embeddings Inference API with a /rerank endpoint.
Parse Score