other
0.57B parameters
Commercial OK
Reviewed June 2026

BGE Reranker v2 M3

BGE M3 reranker. Cross-encoder for re-ranking RAG candidates; multilingual.

License: MIT·Released Apr 15, 2024·Context: 8,192 tokens
BLK · VERDICT

Our verdict

OP · Fredoline Eruo|VERIFIED JUN 12, 2026
unrated

The default open reranker for any production RAG in 2026. Apache-2.0 with 100-language coverage and 8K context is unmatched by any other open reranker at this size. Pair it with bge-m3 or arctic-embed-l-v2 for a complete retrieve-rerank stack.

Overview

BGE M3 reranker. Cross-encoder for re-ranking RAG candidates; multilingual.

Family & lineage

How this model relates to others in its lineage. Family members share architecture and training-data roots; parent / children edges record direct distillation or fine-tune relationships.

Parent / base model
BGE M30.57B
Edge
Family siblings (bge)
BGE Reranker v2 M30.57B
You are here
BGE M30.57B
Edge

Strengths

  • Multilingual 100+ languages with strong cross-lingual reranking (MIRACL nDCG@10 ~70)
  • 8192-token cross-encoder context — full passages rerankable without truncation
  • Apache-2.0, supported natively in TEI, vLLM, sentence-transformers, FastChat
  • Best license/coverage/quality combination among open rerankers in production

Weaknesses

  • 568M cross-encoder is heavy: ~50-100ms per query-doc pair on A10 — batch carefully
  • No GGUF / llama.cpp support — GPU inference is the realistic path
  • FP16 quality drop is non-trivial on long passages; FP32 recommended for production scoring
  • Larger bge-reranker-v2-gemma exists with better English quality if you can pay the 2B param cost

Quantization variants

Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.

QuantizationFile sizeVRAM required
FP161.1 GB2 GB

Get the model

HuggingFace

Original weights

huggingface.co/BAAI/bge-reranker-v2-m3

Source repository — direct quantization required.

Hardware that runs this

Cards with enough VRAM for at least one quantization of BGE Reranker v2 M3.

Compare alternatives

Models worth comparing

Same parameter band, plus what's one tier above and below — so you can decide what actually fits your hardware.

Step down
Smaller — faster, runs on weaker hardware
No verdicted models in the next tier down yet.

Frequently asked

What's the minimum VRAM to run BGE Reranker v2 M3?

2GB of VRAM is enough to run BGE Reranker v2 M3 at the FP16 quantization (file size 1.1 GB). Higher-quality quantizations need more.

Can I use BGE Reranker v2 M3 commercially?

Yes — BGE Reranker v2 M3 ships under the MIT, which permits commercial use. Always read the license text before deployment.

What's the context length of BGE Reranker v2 M3?

BGE Reranker v2 M3 supports a context window of 8,192 tokens (about 8K).

Source: huggingface.co/BAAI/bge-reranker-v2-m3

Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.

Related — keep moving

Alternatives
Before you buy

Verify BGE Reranker v2 M3 runs on your specific hardware before committing money.