olmOCR vs InternVL3.5
Pick olmOCR for bulk PDF to Markdown conversion, especially academic papers with equations and pages mixing print and handwriting. It publishes a score of 82.4 on olmOCR-Bench and a self-hosted cost under $200 per million pages. Pick InternVL3.5 if you need a general vision language model for document question answering, or one model family you can run at sizes from 1B to 241B-A28B.
olmOCRallenai/olmocr | InternVL3.5opengvlab/internvl | |
|---|---|---|
| Type | Open source | Vision model |
| Price | 2 credits / page | 2 credits / page |
| Per 1,000 pages | $2.00 | $2.00 |
| Typical speed | 1.5 s / page | 3.0 s / page |
| Best for | Bulk PDF to text for datasets | Self-hosted document understanding |
| Inputs | PDF, images | images |
| Languages | Mainly English | Multilingual |
| Max pages per request | 1,000 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | No | No |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | No | ✓ Yes |
| 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 olmOCR if
- It scores 82.4 on olmOCR-Bench, above Marker (76.1), MinerU 2.5.4 (75.2) and Mistral OCR (72.0) in the same table.
- Ai2 puts self-hosted cost under $200 per million pages, and a hosted olmOCR-2-7B is available on DeepInfra at $0.09 input / $0.19 output per million tokens.
- It handles equations, tables and handwriting, and strips headers and footers in natural reading order.
Choose InternVL3.5 if
- Nine sizes from 1B to 241B let you trade accuracy for cost, and models up to 30B fit on a single A100.
- It suits document question answering and self-hosted multimodal pipelines, not only PDF conversion.
- The MoE variants 20B-A4B and 30B-A3B activate few parameters per token.
Key differences
| olmOCR | InternVL3.5 | |
|---|---|---|
| What it is | 7B toolkit and model from Ai2 for PDF and image to Markdown | General vision language model family from OpenGVLab, 1B to 241B-A28B |
| Licence | Apache-2.0 (toolkit) | Apache-2.0 |
| How you run it | Self-host with vLLM on a GPU, or use third-party hosted endpoints | Self-host only, with Transformers, LMDeploy or vLLM |
| Price | Under $200 per million pages self-hosted; DeepInfra $0.09 input / $0.19 output per million tokens | Free software; you pay for compute |
| Hardware | NVIDIA GPU with at least 12 GB VRAM and 30 GB free disk | Up to 30B on one A100; 38B needs two A100s; 241B-A28B needs eight A100s |
| Output formats | Markdown, JSON | Plain text, Markdown, JSON |
| OCR benchmark | 82.4 on olmOCR-Bench (v0.4.0), per the authors | Not published as text; the model card shows OCR scores only as chart images |
| Known failure | Out of memory on a 12 GB GPU (issue #424), gets stuck on the last pages (issue #433), unreliable image bounding boxes (issue #437) | No bounding boxes or layout output by default, so no word coordinates |
Sources
What each one is
olmOCR
Open-source vision model from Ai2 (often searched as olm ocr) that turns PDFs into clean, reading-order text. Latest release: olmOCR 2.
olmOCR detailsInternVL3.5
Shanghai AI Lab's open vision-language family (1B to 241B, August 2025), strong on OCR and document benchmarks.
InternVL3.5 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": "allenai/olmocr","fallbacks": ["opengvlab/internvl"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
olmOCR alternativesInternVL3.5 alternativesBest OCR models in 2026