DeepSeek OCR 2
Unavailable in this round
Language model
DeepSeek OCR 2
DeepSeek
- Released
- January 27, 2026
- Data date
- September 26, 2026
- Input
- Text, Image
- Output
- Text
What is distinctive about DeepSeek OCR 2 is its document processing with dynamic resolution. The 3-billion-parameter checkpoint combines multiple crops with an overview and, according to its model card, supports free OCR and document-to-Markdown prompts.
DeepSeek OCR 2 is limited to layouts and text recognition rather than general chat or tool loops. Its model card documents local serving, Apache 2.0, and an 8,192-token context for these OCR pipelines.
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
olmOCR-bench overall (Show measurement, test conditions, and source)
- Source value
- 76.3
- Score
- 76.3
- Metric
- score
- Unit
- No unit provided
- Benchmark version
- olmOCR-bench Overall leaderboard
- Category
- document-ocr
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- Allen Institute olmOCR-bench leaderboard
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Official benchmark leaderboard. Exact DeepSeek-OCR-2 result shown in the official olmOCR-bench leaderboard widget; evaluation setup and run date are not stated in the result. Benchmark prompt, sampling/decoding settings, reasoning effort, quantization, and task run date are not stated in the cited result; none are inferred.
- Context
- The source presents a raw score without an explicit unit or scale; the published value is retained without percentage conversion.
olmOCR-bench Arxiv Math (Show measurement, test conditions, and source)
- Source value
- 82
- Score
- 82
- Metric
- score
- Unit
- No unit provided
- Benchmark version
- olmOCR-bench Arxiv Math subset
- Category
- document-ocr
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- Allen Institute olmOCR-bench leaderboard
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Official benchmark leaderboard. Exact DeepSeek-OCR-2 model-card leaderboard result for the Arxiv Math subset; metric unit and run date are not stated. Benchmark prompt, sampling/decoding settings, reasoning effort, quantization, and task run date are not stated in the cited result; none are inferred.
- Context
- The source presents a raw score without an explicit unit or scale; the published value is retained without percentage conversion.
olmOCR-bench Old Scans Math (Show measurement, test conditions, and source)
- Source value
- 72
- Score
- 72
- Metric
- score
- Unit
- No unit provided
- Benchmark version
- olmOCR-bench Old Scans Math subset
- Category
- document-ocr
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- Allen Institute olmOCR-bench leaderboard
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Official benchmark leaderboard. Exact DeepSeek-OCR-2 model-card leaderboard result for the Old Scans Math subset; metric unit and run date are not stated. Benchmark prompt, sampling/decoding settings, reasoning effort, quantization, and task run date are not stated in the cited result; none are inferred.
- Context
- The source presents a raw score without an explicit unit or scale; the published value is retained without percentage conversion.
Real5-OmniDocBench overall (Show measurement, test conditions, and source)
- Source value
- 73.01
- Score
- 73.01
- Metric
- score
- Unit
- No unit provided
- Benchmark version
- Real5-OmniDocBench Overall
- Category
- document-ocr
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- PaddlePaddle Real5-OmniDocBench leaderboard
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Official benchmark leaderboard. Exact DeepSeek-OCR-2 card evaluation widget for Real5-OmniDocBench overall; prompt/decoding setup and run date are not stated. Benchmark prompt, sampling/decoding settings, reasoning effort, quantization, and task run date are not stated in the cited result; none are inferred.
- Context
- The source presents a raw score without an explicit unit or scale; the published value is retained without percentage conversion.
Our tests
SVG short test
One SVG prompt and fixed output settings. Captures are shown only with 0 reported reasoning tokens or documented lack of reasoning controls. Identical test conditions do not make the models' internal compute identical.
Popularity
42.5 / 100
Sources and data date
| Type | Evidence and data date |
|---|---|
| Input | Documented modalities. Retrieved September 26, 2026. |
| Output | Documented modalities. Retrieved September 26, 2026. |
| Research status | Research date October 15, 2026. 3 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not published | ocr-leaderboard-protocol Note Independent widget results are published, but exact OCR preprocessing, image resolution, decoding settings, and task/run dates are not exposed in the checked source. |
| Additional source | Hugging Face (retrieved October 15, 2026) · DeepSeek OCR 2 model card · Editorial description reviewed October 2, 2026 |
| Additional source | Allen Institute olmOCR-bench leaderboard (retrieved October 15, 2026) |
| Additional source | PaddlePaddle Real5-OmniDocBench leaderboard (retrieved October 15, 2026) |
| Additional source | DeepSeek OCR 2 paper · Editorial description reviewed October 2, 2026 |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Creative Commons Attribution 4.0 |