bge-large-zh-v1.5

bge-large-zh-v1.5 is a BGE-family model for embeddings, reranking, retrieval, semantic search, or RAG workflows.

BAAI · BGE

Editorial review

Reviewed byOpenSourcesAI EditorialLast updatedAugust 2026SourcesExact model card

Model checkpoints, context windows, provider support, local runtime compatibility, and license terms can change quickly. Verify the exact model card before production or commercial use.

Best for

Chinese embedding and retrieval workflows

Who should use it

Chinese embedding and retrieval workflows
Builders who want local or self-hosted testing options.
RAG builders comparing retrieval quality, latency, and multilingual coverage.

Common workflows

Chinese embedding and retrieval workflows
embedding workflows
rag workflows
retrieval workflows
semantic search workflows

Deployment and hardware notes

~1.30 GB in fp32 (326M parameters), ~0.65 GB in fp16 — CPU works for prototypes, and any modern GPU serves it interactively.

License and usage notes

MIT. Open weights where released. Verify the exact model card and license terms for the checkpoint or hosted provider you use.

Strengths

Open weights where released model option for BGE workflows.
Chinese embedding and retrieval workflows
Chinese queries use the model card's own instruction prefix (为这个句子生成表示以用于检索相关文章：); passages are embedded bare.

Limitations

Chinese-first: it is the wrong pick for English or mixed corpora, where bge-m3 or the English v1.5 line will rank better. Inputs cap at 512 tokens.
~1.30 GB in fp32 (326M parameters), ~0.65 GB in fp16 — CPU works for prototypes, and any modern GPU serves it interactively.
Context window and limits: 512 tokens.
Verify the exact model card, provider docs, license, and serving support before production use.

Local workflow notes

Chinese queries use the model card's own instruction prefix (为这个句子生成表示以用于检索相关文章：); passages are embedded bare.

Local runtimes: Ollama where supported, LM Studio where supported, llama.cpp where supported, Transformers

Platforms: Windows, macOS, Linux

Embedding spec

Memory~1.30 GB in fp32 (326M parameters) · ~0.65 GB in fp16Output1024-dimension vectorsMax input512 tokens

BAAI's large Chinese embedder: 1024-dimension vectors tuned for Chinese retrieval benchmarks. Pair it with bge-reranker-base or bge-reranker-large, which were trained on Chinese and English pairs.