NEWVectors or files. Pick a path.Start →
    Models/Reranking/LiquidAI/LFM2.5-ColBERT-350M
    HFRerankingother

    LFM2.5-ColBERT-350M

    by LiquidAI

    Edge-sized ColBERT late-interaction retriever built on LFM2.5

    12Kdl/month
    107likes
    353Mparams
    Identifiers
    Model ID
    LiquidAI/LFM2.5-ColBERT-350M
    Feature URI
    mixpeek://reranker@v1/lfm25_colbert_350m_v1

    Overview

    LFM2.5-ColBERT-350M is Liquid AI's compact late-interaction retriever: instead of collapsing a document into one vector, it keeps a small vector per token and scores relevance with MaxSim between query and document token vectors. That preserves term-level precision that single-vector embeddings blur away, which shows up on exact-phrase, entity-heavy, and long-document retrieval. At 350M parameters with the LFM2.5 backbone's efficiency, it runs late-interaction quality at edge and on-CPU budgets where classic ColBERT deployments were impractical.

    On Mixpeek, a late-interaction model like this slots into the reranking or precision stage of a multi-stage retriever: a dense first stage recalls candidates cheaply, then token-level MaxSim re-scores the top-K. See the late interaction retrieval guide for when token-level matching beats single vectors and what it costs at the index layer.

    Architecture

    ColBERT-style late-interaction architecture (PyLate-compatible) on the LFM2.5 hybrid backbone: per-token contextual embeddings with MaxSim scoring, trained for sentence-similarity and retrieval. 350M parameters, English-focused, runs via sentence-transformers/PyLate with custom code enabled.

    Mixpeek SDK Integration

    // Index documents in a collection (dense first stage)
    const collection = await mx.collections.create({
      collection_name: "docs-collection",
      source: { type: "bucket", bucket_ids: ["bkt_your_bucket"] },
      feature_extractor: {
        feature_extractor_name: "text_embeddings",
        version: "v1",
        parameters: { model_id: "BAAI/bge-m3" },
      },
    });

    Capabilities

    • Token-level late-interaction scoring (MaxSim) for precise term matching
    • Multi-vector document representations that survive exact-phrase queries
    • Edge/CPU-friendly footprint at 350M parameters
    • PyLate and sentence-transformers compatible for drop-in retrieval stacks
    • Strong fit as a precision reranking stage over a dense first stage

    Use Cases on Mixpeek

    Reranking top-K candidates from a dense retriever with token-level precision
    Entity- and phrase-heavy corpora (legal, technical docs) where single vectors blur terms
    On-prem or edge retrieval where large rerankers do not fit
    Hybrid stacks pairing BM25 or dense recall with late-interaction re-scoring

    Benchmarks

    DatasetMetricScoreSource
    Hugging FacePipeline tagsentence-similarity (ColBERT/PyLate)Liquid AI, 2026 Model Card
    Hugging FaceParameters350MLiquid AI, 2026 Model Card
    Hugging FaceBackboneLFM2.5Liquid AI, 2026 Model Card

    Performance

    Input SizeText queries and documents (per-token vectors)
    GPU LatencyMaxSim scoring scales with candidate count and doc length
    GPU ThroughputUse as a rerank stage after top-K candidate generation
    GPU MemoryEdge-friendly at 350M parameters; multi-vector index grows with token count

    Specification

    FrameworkHF
    OrganizationLiquidAI
    FeatureReranking
    OutputRelevance score per candidate
    Modalities
    RetrieverCross-Modal Reranker
    Parameters353M
    Licenseother
    Downloads/mo12K
    Likes107

    Research Paper

    LFM2.5-ColBERT-350M model card

    arxiv.org

    Build a pipeline with LFM2.5-ColBERT-350M

    Add this model to a processing pipeline alongside other extractors. Combine with retrieval stages for end-to-end search.

    Run it on your own data, free