# ByteDance-Seed/Seed1.5-VL

Seed1.5-VL, a vision-language foundation model designed to advance general-purpose multimodal understanding and reasoning, achieving state-of-the-art performance on 38 out of 60 public benchmarks.

Repository: https://github.com/ByteDance-Seed/Seed1.5-VL
Canonical: https://ross.abutalabs.com/products/seed15-vl
Homepage: https://seed.bytedance.com/en/tech/seed1_5_vl
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Topics: cookbook, large-language-model, multimodal-large-language-models, vision-language-model
Last push: 2025-06-14T19:58:52+00:00

## Health v2 (maintenance only)
Score: 31/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 26, release rhythm 35, longevity 34
- inputs: {"age_days": 479, "days_push": 445, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1586, forks 66 (observed 2026-08-28T04:05:07.718833+00:00)

## What it is
Seed1.5-VL is ByteDance's vision-language foundation model (532M vision encoder + 20B active-parameter MoE LLM) for general-purpose multimodal understanding and reasoning, state-of-the-art on 38 of 60 public benchmarks. This repository is the official cookbook with code samples and best practices for using the model via its hosted API on Volcano Engine.

## Use cases
- understand images and diagrams with a vision-language model
- extract text from images with ocr
- build gui agents that control apps from screenshots
- analyze video content with a multimodal model
- ground and locate objects in 2d and 3d scenes
- solve visual reasoning puzzles with an ai model

## When to choose
- you want to integrate a top-performing vision-language model via API
- you need cookbook examples for grounding, video understanding, or GUI agents
- you want to evaluate a strong multimodal model against benchmarks

## When to avoid
- you need to run the model weights locally - the repo only provides API usage examples
- you want a self-hosted or fine-tunable open-weights model
- you need a model without dependency on Volcano Engine's paid API

## Facets
- artifact type: learning-resource
- maturity: active
- function: machine-learning, nlp, computer-vision, ocr, sdk
- domain: large-language-models, computer-vision, artificial-intelligence, tutorials
- platform: python, cloud
- tags: vision-language-model, multimodal, cookbook, moe, foundation-model, gui-agents, video-understanding, visual-grounding, ai-agents, web-server

## Member repositories
- ByteDance-Seed/Seed1.5-VL (main) score 31

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:07.718833+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:55:26.955468+00:00, confidence not recorded.
  - readme: https://github.com/ByteDance-Seed/Seed1.5-VL (fetched 2026-08-28T04:05:07.718833+00:00, sha d54d298ba9ea)
  - homepage: https://seed.bytedance.com/en/tech/seed1_5_vl (fetched 2026-08-29T11:26:04.560753+00:00, sha 605f70939a47)
- Data as of 2026-08-30T08:39:29.467469+00:00.