Jina Embeddings v5 Text Small Text-Matching by Jina AI — 1024D

677M-parameter text-matching-targeted variant of v5-text-small. Optimized for symmetric pairwise similarity scoring, STS, paraphrase, and near-duplicate detection. 1024-dim embeddings with Matryoshka truncation. Supports 119+ languages up to 32K tokens. Available in GGUF, ONNX, and BF16 formats. Compatible with vLLM, TEI, llama.cpp, and sentence-transformers.