Janus Models by DeepSeek
1 model2025
Last refreshed 2026-05-09. Next refresh: weekly.
Details
ResearcherDeepSeek
LicenseMITOSI-approved
Commercial useCommercial use: permitted
Models1
Released2025
Capabilities
VisionAll models
MultimodalAll models
Links
WebsiteAbout
DeepSeek's Janus family of multimodal models supporting both image understanding and image generation. Open-weight models bridging vision understanding and generation in a unified architecture.
Decision facts
- Best fit
- image generationvision and multimodal work
- Capability starting point
- Janus Pro with multimodal inputs
- Lowest tracked input
- Not tracked
- Closest related family
- DeepSeek OCR
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
1 in view
Janus ProCurrent
Use when the workload needs image generation and multimodal inputs.
2025-01image generationmultimodal inputs
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Janus Pro | Use when the workload needs image generation and multimodal inputs. | 2025-01 | image generationmultimodal inputs | Current |
Release Timeline
1 release group2025-01
1 current
Janus Pro
Currentimage generationmultimodal inputs
Specifications(1 models)
| Model | Released | Vision | Multimodal |
|---|---|---|---|
| Janus Pro | 2025-01 | Yes | Yes |
