# deepseek-ai/DeepSeek-VL2

DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding

Repository: https://github.com/deepseek-ai/DeepSeek-VL2
Canonical: https://ross.abutalabs.com/products/deepseek-vl2
Language: Python
License: MIT
License Family: permissive
Last push: 2025-02-26T05:03:42+00:00

## Health v2 (maintenance only)
Score: 25/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 8, release rhythm 35, longevity 44
- inputs: {"age_days": 628, "days_push": 553, "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 5374, forks 1810 (observed 2026-08-28T04:09:16.180387+00:00)

## What it is
DeepSeek-VL2 is a series of Mixture-of-Experts vision-language models (Tiny, Small, and 4.5B activated parameters) with inference code and model weights for advanced multimodal understanding. It handles tasks like visual question answering, OCR, document/table/chart understanding, and visual grounding.

## Use cases
- run a vision-language model for image question answering
- extract text from images with OCR
- understand documents, tables, and charts from screenshots
- ground and locate objects in images from natural language queries
- build multimodal chat applications that see images
- compare MoE vision-language model variants by size

## When to choose
- you need strong multimodal understanding with efficient MoE inference
- you want an open-weights VLM with MIT-licensed code
- you need OCR or document/chart understanding out of the box

## When to avoid
- you need a text-only LLM
- you cannot run GPU inference locally
- you need a permissively licensed model for commercial use without checking the model agreement

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, llm-inference, ocr, image-processing, nlp
- domain: artificial-intelligence, large-language-models, computer-vision, deep-learning
- platform: python
- tags: vision-language-model, mixture-of-experts, multimodal, visual-question-answering, visual-grounding, model-weights, natural-language-processing, gpu, linux, docker

## Member repositories
- deepseek-ai/DeepSeek-VL2 (main) score 25

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:16.180387+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 2026-08-29T17:58:41.864662+00:00, confidence not recorded.
  - readme: https://github.com/deepseek-ai/DeepSeek-VL2 (fetched 2026-08-28T04:09:16.180387+00:00, sha 833a5cb30468)
- Data as of 2026-08-30T08:39:29.467469+00:00.