Qianfan OCR vs InternVL3.5
Pick Qianfan OCR when you need a ready-made document parser that returns Markdown, HTML, LaTeX or bounding boxes, with a published OmniDocBench v1.5 score of 93.12. Pick InternVL3.5 when you want a general vision language model for document question answering and need to choose a size from 1B to 241B. Both are Apache-2.0, but only Qianfan OCR publishes OCR scores as numbers and gives layout coordinates.
Qianfan OCRbaidu/qianfan-ocr | InternVL3.5opengvlab/internvl | |
|---|---|---|
| Type | OCR API | Vision model |
| 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 | Chinese documents | Self-hosted document understanding |
| Inputs | PDF, images | images |
| Languages | Chinese, English and more | Multilingual |
| Max pages per request | 200 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | No | No |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | ✓ Yes | ✓ 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 Qianfan OCR if
- It scores 93.12 on OmniDocBench v1.5, 79.8 on olmOCR-Bench and 880 on OCRBench, all published as numbers on the model card.
- It outputs Markdown, JSON, HTML, LaTeX and bounding boxes, so you can get layout and coordinates without extra tooling.
- It covers 192 languages, handwriting and scene text, and it is also offered hosted on the Baidu Qianfan platform if you do not want to run it yourself.
Choose InternVL3.5 if
- It comes in nine sizes from 1B to 241B-A28B, so you can trade accuracy for cost within one family.
- Models up to 30B run on a single A100, and the MoE variants 20B-A4B and 30B-A3B activate few parameters per token.
- It suits document question answering and self-hosted multimodal pipelines, where OCR is one task among several.
Key differences
| Qianfan OCR | InternVL3.5 | |
|---|---|---|
| Model type | 4B end-to-end document parsing model (Qwen3-4B language model) | General vision language model family, 1B to 241B-A28B |
| Licence | Apache-2.0, commercial use allowed | Apache-2.0, built on Apache-2.0 Qwen3 |
| How you run it | Self-host (benchmarked on one A100 with vLLM) or use the Baidu Qianfan platform | Self-host only, with Transformers, LMDeploy or vLLM |
| Output formats | Markdown, JSON, HTML, LaTeX, bounding boxes | Plain text, Markdown, JSON. No bounding boxes or layout output by default |
| Published OCR benchmarks | OmniDocBench v1.5 93.12, olmOCR-Bench 79.8, OCRBench 880 (author-reported) | Not published as text values; the model card shows OCR scores only as chart images |
| Hardware | One A100 with vLLM in the benchmark; no minimum VRAM stated | Up to 30B on one A100, 38B on two A100s, 241B-A28B on eight A100s |
| Known weakness | 1.024 pages per second on an A100 even with W8A8, and thinking mode can emit up to 16,384 tokens | Not suited to jobs needing word coordinates; larger sizes are hard to run on CPU only |
Sources
What each one is
Qianfan OCR
Baidu's OCR-tuned vision model. Strong on Chinese documents and tables.
Qianfan OCR 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": "baidu/qianfan-ocr","fallbacks": ["opengvlab/internvl"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Qianfan OCR alternativesInternVL3.5 alternativesBest OCR models in 2026