BBrainOutput
BAAI·Embedding·MIT·BAAI·2024

BGE-M3 Embeddings (class): Hardware & Business Fit

  • Multilingual
  • Embedding

A multilingual embedding model for retrieval. Embeddings power the search step in RAG; they run alongside, not instead of, your chat model.

Parameters
~0.6B
Context
~8K tokens
Deployment
local

What BGE-M3 Embeddings (class) is good for

  • Multilingual RAG
  • Long-document retrieval
  • Legal/finance search
multilingual retrievallong documentsRAG

Best quantization choices

Approximate memory per quantization (weights + KV cache at modest context). Treat as ±.

Quant~MemoryWhen to use
FP16~2GBFull precision; largest footprint, best quality.

Run BGE-M3 Embeddings (class) locally

Pull and run with Ollama, or grab the weights from Hugging Face.

$ ollama run bge-m3
Hugging Face repo
BAAI/bge-m3

Compatible hardware

Devices from our catalog graded for BGE-M3 Embeddings (class), best fit first.

  • NVIDIA B200 (placeholder)
    NVIDIA · Datacenter GPUs

    Fits at FP16 (~2GB) with ~167GB headroom — about 84 concurrent instances.

    FP16 · ~2GBRuns well
  • Supermicro 8x H100 SuperServer
    Supermicro · AI Servers

    Fits at FP16 (~2GB) with ~561.2GB headroom — about 281 concurrent instances.

    FP16 · ~2GBRuns well
  • Dell PowerEdge XE9680
    Dell · AI Servers

    Fits at FP16 (~2GB) with ~561.2GB headroom — about 281 concurrent instances.

    FP16 · ~2GBRuns well
  • AMD Instinct MI300X
    AMD · Datacenter GPUs

    Fits at FP16 (~2GB) with ~167GB headroom — about 84 concurrent instances.

    FP16 · ~2GBRuns well
  • Cloud B200 (Blackwell profile, to verify)
    Cloud · Cloud GPU Profiles

    Fits at FP16 (~2GB) with ~156.4GB headroom — about 79 concurrent instances.

    FP16 · ~2GBRuns well
  • NVIDIA H200 (141GB)
    NVIDIA · Datacenter GPUs

    Fits at FP16 (~2GB) with ~122.1GB headroom — about 62 concurrent instances.

    FP16 · ~2GBRuns well
  • Cloud H200 141GB (profile)
    Cloud · Cloud GPU Profiles

    Fits at FP16 (~2GB) with ~122.1GB headroom — about 62 concurrent instances.

    FP16 · ~2GBRuns well
  • NVIDIA H100 (80GB)
    NVIDIA · Datacenter GPUs

    Fits at FP16 (~2GB) with ~68.4GB headroom — about 35 concurrent instances.

    FP16 · ~2GBRuns well
  • Cloud H100 80GB (profile)
    Cloud · Cloud GPU Profiles

    Fits at FP16 (~2GB) with ~68.4GB headroom — about 35 concurrent instances.

    FP16 · ~2GBRuns well
  • NVIDIA RTX PRO 6000 Blackwell
    NVIDIA · Professional GPUs

    Fits at FP16 (~2GB) with ~82.5GB headroom — about 42 concurrent instances.

    FP16 · ~2GBRuns well

Use inside the AI Business OS

BGE-M3 Embeddings (class) suits these AI Business OS agent archetypes:

A model is only the engine. Inside the AI Business OS it is wrapped with permissions, tools, connectors, RAG and audit so it can actually do business work safely — see how the AI Business OS works →

Frequently asked questions

What hardware do I need to run BGE-M3 Embeddings (class)?+

At 4-bit you need roughly a few GB of usable memory. The minimum self-hostable option in our catalog is the NVIDIA GeForce RTX 3060 12GB. For a comfortable run we recommend the NVIDIA B200 (placeholder).

Which quantization should I use for BGE-M3 Embeddings (class)?+

Q4_K_M is the usual default — the best size/quality trade-off. Step up to Q8_0 or FP16 if you have spare memory and want higher fidelity.

Should I run BGE-M3 Embeddings (class) locally or in the cloud?+

Local-first is recommended for BGE-M3 Embeddings (class). It fits comfortably on hardware you can own, keeping data private and costs predictable.

Related models

Similar picks — family siblings and nearest-size models of the same kind.

Use BGE-M3 Embeddings (class) inside your AI Business OS

BrainOutput helps you run BGE-M3 Embeddings (class) as a private business agent — wrapped with the tools, connectors, RAG and guardrails it needs to do real work on hardware you control.

Use this model in your AI Business OS