Chandra vs RolmOCR
Pick Chandra when you need structured output (Markdown, HTML, JSON with bounding boxes), tables, forms, math or non-English scans, and you accept its OpenRAIL-M licence terms or the paid Datalab API. Pick RolmOCR when you want plain text from scanned pages under a plain Apache-2.0 licence on your own GPU, and can tolerate some dropped lines. The main split is output: Chandra returns layout and coordinates, RolmOCR returns plain text only.
Chandradatalab/chandra | RolmOCRreducto/rolmocr | |
|---|---|---|
| Type | Open source | Open source |
| Price | 2 credits / page | 2 credits / page |
| Per 1,000 pages | $2.00 | $2.00 |
| Typical speed | 2.0 s / page | 3.0 s / page |
| Best for | Complex layouts and handwriting | Fast open PDF to text |
| Inputs | PDF, images | PDF, images |
| Languages | 40+ languages | Not stated by the authors |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | ✓ Yes |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | No |
| Confidence | No | No |
| Forms | ✓ Yes | No |
| Multilingual | ✓ Yes | No |
| On OCRRouter | Available | Available |
Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.
Which to choose
Choose Chandra if
- It outputs Markdown, HTML and JSON plus layout chunks with bounding boxes and labels per block, which RolmOCR cannot do.
- Chandra 2 scores 85.8 on olmOCR-bench and averages 77.8% across 43 languages, so it fits multilingual scans.
- You can skip GPU hosting and use the Datalab API at $1, $3 or $7.50 per 1,000 pages, with a free tier of $20/month.
Choose RolmOCR if
- Apache-2.0 weights have no revenue or funding cap, so any company can self-host them commercially for free.
- About 15% of its training data was rotated, which suits off-angle and skewed scans.
- Needl.ai counted 7 hallucination issues for RolmOCR versus 11 for olmocr-2-7b, and Reducto reports faster inference and lower memory use than the original olmOCR.
Key differences
| Chandra | RolmOCR | |
|---|---|---|
| Licence | Code Apache-2.0; weights modified OpenRAIL-M, free under $2M funding/revenue, commercial licence needed above that | Apache-2.0, usable commercially |
| How you run it | Self-host with vLLM or Hugging Face, or use the hosted Datalab API | Self-host only, for example with vLLM |
| Price | Self-hosted: compute only within licence terms. API: $1, $3 or $7.50 per 1,000 pages | Free, you pay for compute |
| Output formats | Markdown, HTML, JSON, bounding boxes | Plain text; no layout bounding boxes |
| Benchmark | 85.8 on olmOCR-bench (Chandra 2, authors' figure) | 0.663 overall on Needl.ai (independent), versus 0.770 for olmocr-2-7b |
| Hardware | Single NVIDIA GPU via vLLM; benchmarked on one H100 80GB, presets for A100, L40S, RTX 4090/5090 | Self-hosted on one GPU; 8B params (Qwen2.5-VL-7B base) |
| Known failure | Token-loop hallucinations on 5-6 of about 60 Russian/Kazakh scans (GitHub issue #51); weak on Urdu (43.2%) and old scans (51.1) | Model card says it can hallucinate or drop content; omitted a table subtitle in Reducto's example |
Sources
- Chandra GitHub README
- Datalab pricing
- GitHub issue #51 token loops
- Hugging Face model card
- Reducto announcement
- Needl.ai benchmark
What each one is
Chandra
Datalab's OCR model that keeps layout and handles forms, tables and handwriting.
Chandra detailsRolmOCR
Reducto's open PDF-to-text model, built on Qwen2.5-VL-7B with the olmOCR recipe. Needs no PDF metadata and handles rotated pages.
RolmOCR detailsTry both on your own file
Run the file once with each model and compare the output.
Use one, fall back to the other
{"model": "datalab/chandra","fallbacks": ["reducto/rolmocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Chandra alternativesRolmOCR alternativesBest OCR models in 2026