Lance

Name: Lance
Author: bytedance-research

by bytedance-research

Unified 3B model for image and video understanding, generation, and editing

32Kdl/month

3Bparams

HuggingFace Run it on your own data, free

Identifiers

Model ID

bytedance-research/Lance

Feature URI

mixpeek://video_extractor@v1/bytedance_lance_3b_v1

Overview

Lance is ByteDance's 3B-parameter unified vision model that handles image understanding, video understanding, image generation, video generation, and image/video editing in a single architecture. It uses a vision tokenizer to convert between continuous pixel space and discrete token space, enabling a shared transformer to reason across both modalities.

On Mixpeek, Lance is relevant as a compact video understanding model that can caption, describe, and answer questions about both images and video content. Its unified architecture means a single model can power scene description, visual Q&A, and content analysis pipelines.

Architecture

Unified autoregressive transformer with a learned vision tokenizer. 3B parameters. Supports text-to-image, text-to-video, image/video understanding, and editing through a shared token space.

Mixpeek SDK Integration

// No extractor parameter takes a Hugging Face model id (checked against
// GET /v1/discovery/extractors, which returns 13), so Lance runs
// on your side and the output is upserted through POST
// /v1/namespaces/{namespace_id}/documents/upsert. On Enterprise the other
// path is to upload the weights instead: POST /v1/namespaces/{id}/models
// accepts the huggingface format and a custom plugin loads them.
const res = await fetch(
  "https://api.mixpeek.com/v1/namespaces/ns_your_namespace/documents/upsert",
  {
    method: "POST",
    headers: {
      Authorization: "Bearer API_KEY",
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      collection_id: "col_your_collection",
      documents: [
        {
          document_id: "asset-00412",
          // The model produces text, so it lands in payload. Give the
          // collection a text vector index and embed that text to make it
          // searchable rather than only filterable.
          payload: { extracted_text: modelOutput, source_key: "archive/2026/asset-00412" },
          vectors: { "multimodal-embedding": embeddingOfModelOutput },
        },
      ],
    }),
  },
);

// Managed alternative, if this exact model is not the requirement:
// universal_extractor@v1 runs google/gemini-embedding-2
// (3072-d) over a bucket, with no inference of your own.