OCRRouter

Claude vs InternVL3.5

Pick Claude to read long PDFs, answer questions with page citations, and pull structured JSON out of reports through a hosted API. Pick InternVL3.5 when you must self-host, want an Apache-2.0 licence, or need to choose a model size from 1B to 241B. Claude costs 1,500 to 3,000 tokens per PDF page, so bulk OCR is cheaper on a model you run yourself.

Claudeanthropic/claudeInternVL3.5opengvlab/internvl
TypeVision modelVision model
Price25 credits / page2 credits / page
Per 1,000 pages$25.00$2.00
Typical speed2.6 s / page3.0 s / page
Best forLong dense documentsSelf-hosted document understanding
InputsPDF, imagesimages
Languages50+ languagesMultilingual
Max pages per request100500
Tables✓ Yes✓ Yes
Handwriting✓ YesNo
Math✓ Yes✓ Yes
Bounding boxesNoNo
ConfidenceNoNo
FormsNoNo
Multilingual✓ Yes✓ Yes
On OCRRouterAvailableAvailable

Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.

Which to choose

Choose Claude if

  • It takes PDFs natively and gives page-level citations, up to 600 pages per request on 1M-context models (100 pages otherwise) with a 32 MB payload.
  • It reads each page as both extracted text and a rendered image, so charts and scans are visible to the model.
  • It runs as a hosted API with no GPUs to manage, and it can return Markdown, JSON or plain text.

Choose InternVL3.5 if

  • It is free to self-host under Apache-2.0, so you pay only for compute.
  • Nine sizes from 1B to 241B let you trade accuracy for cost, and models up to 30B fit on a single A100.
  • The MoE variants 20B-A4B and 30B-A3B activate few parameters per token.

Key differences

ClaudeInternVL3.5
How you run itHosted APISelf-host with Transformers, LMDeploy or vLLM
LicenceProprietaryApache-2.0
PriceToken-based, no extra PDF fee. A PDF page typically uses 1,500 to 3,000 tokens plus image tokens.Free software. You pay for compute.
HardwareNone needed on your sideUp to 30B on one A100, 38B on two A100s, 241B-A28B on eight A100s
Input limits600 pages per request on 1M-context models, 100 otherwise, 32 MB payload. Password-protected or encrypted PDFs are not supported.Not published
Bounding boxesNot suited to pipelines that need bounding boxesNo bounding boxes or layout output by default, and not suited to jobs needing word coordinates
Published OCR scoresNo published OmniDocBench score for any Claude model in the leaderboards checkedModel card shows OCR scores only as chart images, not text values

Sources

What each one is

Claude

General vision model with careful reading of long, dense documents.

Claude details

InternVL3.5

Shanghai AI Lab's open vision-language family (1B to 241B, August 2025), strong on OCR and document benchmarks.

InternVL3.5 details

Try both on your own file

Run the file once with each model and compare the output.

PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?

Use one, fall back to the other

{"model": "anthropic/claude","fallbacks": ["opengvlab/internvl"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

Claude alternativesInternVL3.5 alternativesBest OCR models in 2026