# deepseek-ai/Janus

Janus-Series: Unified Multimodal Understanding and Generation Models

Repository: https://github.com/deepseek-ai/Janus
Canonical: https://ross.abutalabs.com/products/janus
Language: Python
License: MIT
License Family: permissive
Topics: any-to-any, foundation-models, llm, multimodal, vision-language-pretraining, unified-model
Last push: 2025-02-01T07:58:29+00:00

## Health v2 (maintenance only)
Score: 24/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 4, release rhythm 35, longevity 48
- inputs: {"age_days": 684, "days_push": 578, "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 17757, forks 2237 (observed 2026-08-28T04:11:20.335711+00:00)

## What it is
Janus-Series is DeepSeek's family of unified multimodal models (Janus, Janus-Pro, JanusFlow) that combine multimodal understanding and image generation in a single model. The repository provides model weights, inference code, and demos for running these models.

## Use cases
- generate images from text prompts
- answer questions about images with a vision-language model
- run a unified multimodal understanding and generation model locally
- fine-tune or evaluate a multimodal foundation model
- compare Janus-Pro and JanusFlow model variants

## When to choose
- you need one model for both image understanding and text-to-image generation
- you want open weights under a permissive code license for research
- you want a lightweight (1.3B) multimodal model for demos

## When to avoid
- you need production-grade image generation quality on par with dedicated diffusion models
- you cannot run GPU inference
- you need a fully managed API rather than self-hosted model weights

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, llm-inference, image-processing, nlp
- domain: artificial-intelligence, large-language-models, computer-vision, image-processing
- platform: python
- tags: multimodal, any-to-any, vision-language-model, text-to-image, image-generation, foundation-models, deepseek, natural-language-processing, gpu, linux

## Member repositories
- deepseek-ai/Janus (main) score 24

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:20.335711+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 2026-08-29T17:02:28.459525+00:00, confidence not recorded.
  - readme: https://github.com/deepseek-ai/Janus (fetched 2026-08-28T04:11:20.335711+00:00, sha 805f13474fc9)
- Data as of 2026-08-30T08:39:29.467469+00:00.