mxbai-rerank

Open weights · mixedbread-ai · Updated 21 Jun 2026

mixedbread-ai's reranker family is built on DeBERTa-v3 — an architecture with disentangled attention that delivers strong cross-encoder precision. All three variants ship under Apache 2.0, so you can self-host freely. The xsmall model is compact enough to run in the browser, which is why it's one of the options in this site's live demo.

Model variants

All three are DeBERTa-v3 cross-encoders trained on MS MARCO passage ranking. Start with mxbai-rerank-base-v1 for production use; switch to large if you have GPU headroom and need the extra precision. Use xsmall for browser or edge deployments where model size is the constraint.

Benchmarks

Scores are approximate averages across the 18 BEIR datasets. Check the mixedbread-ai HuggingFace model cards for per-dataset results.

Quick start

Self-hosted (sentence-transformers)

pip install sentence-transformers

from sentence_transformers import CrossEncoder

model = CrossEncoder("mixedbread-ai/mxbai-rerank-base-v1")

query = "How do I add reranking to my RAG pipeline?"
passages = [
    "Rerankers score each query-passage pair with a cross-encoder.",
    "BM25 is a classical keyword-based retrieval method.",
    "London is the capital of the United Kingdom.",
    "Two-stage retrieval: retrieve 50 candidates, rerank to top 5.",
]

scores = model.predict([(query, p) for p in passages])
ranked = sorted(zip(scores, passages), reverse=True)

for score, text in ranked:
    print(f"{score:.4f}  {text[:80]}")

In a RAG pipeline

from sentence_transformers import CrossEncoder

reranker = CrossEncoder("mixedbread-ai/mxbai-rerank-large-v1")

def rag_answer(query: str, vector_db, llm) -> str:
    # Stage 1: retrieve wide
    candidates = vector_db.search(query, top_k=50)
    # Stage 2: rerank tight
    scores = reranker.predict([(query, c) for c in candidates])
    top5 = [c for _, c in sorted(zip(scores, candidates), reverse=True)[:5]]
    # Stage 3: generate
    return llm.complete(f"Context:\n" + "\n\n".join(top5) + f"\n\nQ: {query}")

Hosted API

mixedbread-ai offers a hosted rerank endpoint backed by the same model weights. Useful if you want to avoid running inference on your own infrastructure.

pip install mixedbread-ai

from mixedbread_ai import MixedbreadAI

mxbai = MixedbreadAI(api_key="YOUR_API_KEY")

result = mxbai.reranking(
    model="mixedbread-ai/mxbai-rerank-large-v1",
    query="How do I add reranking to my RAG pipeline?",
    input=[
        "Rerankers score each query-passage pair jointly.",
        "BM25 is a keyword-based retrieval method.",
        "London is the capital of the United Kingdom.",
        "Two-stage retrieval: retrieve wide, rerank tight.",
    ],
    top_k=3,
    return_input=False,
)

for item in result.data:
    print(f"{item.score:.4f}  rank {item.index + 1}")

Browser use

The xsmall variant is compact enough to run in the browser via transformers.js — this is exactly what our demo uses:

// transformers.js (ES module)
import { AutoTokenizer, AutoModelForSequenceClassification }
  from "https://cdn.jsdelivr.net/npm/@huggingface/transformers@3";

const tokenizer = await AutoTokenizer.from_pretrained(
  "mixedbread-ai/mxbai-rerank-xsmall-v1", { dtype: "q8" }
);
const model = await AutoModelForSequenceClassification.from_pretrained(
  "mixedbread-ai/mxbai-rerank-xsmall-v1", { dtype: "q8" }
);

const inputs = tokenizer(
  [query, query],
  { text_pair: [doc1, doc2], padding: true, truncation: true }
);
const { logits } = await model(inputs);
const scores = logits.sigmoid().tolist();

With dtype: "q8" the quantized weights are roughly 35 MB — fast to download and cached in IndexedDB after the first run.

Pros and cons

Other models