DeepSeek OCR 2
Released
2026-01-28
Last refreshed
2026-06-29
Status
Researched 105d ago
Open sourceCommercial use: permittedMultimodalVision
DeepSeek OCR 2 is worth evaluating for vision when its provider route and context window match the workload.
Use it for
- Teams evaluating vision
- Workloads that can use a 8k context window
- Buyers comparing 1 tracked provider route
Do not use it for
- Strict JSON or tool-calling flows
Specifications
- Family
- DeepSeek OCR
- Released
- 2026-01-28
- Context
- 8k
- Architecture
- Decoder Only
- Specialization
- vision
- Openness
- Open source
- License
- MITOSI-approvedCommercial use: permitted
- Weights
- Available
- Code
- Unknown
- Training
- Pretrained
Created by
Pricing
Output / 1M
$0.030
Input / 1M
$0.030
Cheapest of 1 route · Novita AI
Links
About
DeepSeek OCR 2 is the second-generation OCR vision-language model from DeepSeek, introducing Visual Causal Flow as a key innovation for improved document understanding and recognition. Released with arXiv paper 2601.20552 (January 2026). 1.5M+ downloads per month on HuggingFace.
Top use-case fit
Vision
Included by capability and metadata signals in the decision map.
Provider price ladder
Compare API pricing across 1 providers for input and output tokens, batch, and cached reads when available.
| Provider | Input / 1M | Output / 1M | Route |
|---|---|---|---|
| Novita AI | $0.030 | $0.030 | Serverless |
Capabilities
VisionMultimodal
Benchmark peer barsfor Vision
No task-mapped benchmark peers are available for this model yet.
Migration checks
No linked migration route is available for this model yet.
Created by
Pricing
Output / 1M
$0.030
Input / 1M
$0.030
Cheapest of 1 route · Novita AI
Links