OCRRouter

Gemini Flash vs Chandra

Pick Gemini Flash for cheap bulk PDF to Markdown or JSON extraction when a hosted API is fine. Pick Chandra when you need to self-host, want bounding boxes, or process multilingual scans. The main split is deployment: Gemini Flash is API-only and billed by token, while Chandra has open weights you run on your own GPU.

Gemini Flashgoogle/gemini-flashChandradatalab/chandra
TypeVision modelOpen source
Price15 credits / page2 credits / page
Per 1,000 pages$15.00$2.00
Typical speed1.6 s / page2.0 s / page
Best forHandwriting and mixed layoutsComplex layouts and handwriting
InputsPDF, imagesPDF, images
Languages100+ languages40+ languages
Max pages per request1,000500
Tables✓ Yes✓ Yes
Handwriting✓ Yes✓ Yes
Math✓ Yes✓ Yes
Bounding boxesNo✓ Yes
ConfidenceNoNo
FormsNo✓ Yes
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 Gemini Flash if

  • You want a hosted API with no GPU to run, through the Gemini API or Vertex AI.
  • You need to extract JSON from documents with a prompt or schema, or convert PDFs to Markdown in bulk at low cost.
  • You have long documents, since the context window goes up to 1M tokens and fits them in one request.

Choose Chandra if

  • You must self-host or run on-prem, since the weights run on a single NVIDIA GPU via vLLM.
  • You need layout chunks with bounding boxes and labels per block, plus Markdown, HTML or JSON.
  • You process multilingual scans, where Chandra 2 averages 77.8% across 43 languages versus 67.6% for Gemini 2.5 Flash.

Key differences

Gemini FlashChandra
How you run itHosted API only (Gemini API or Vertex AI)Self-host the open weights (vLLM or Hugging Face) or use the hosted Datalab API
LicenceProprietaryCode Apache-2.0; weights modified OpenRAIL-M, free for research, personal use and companies under $2M funding/revenue, commercial licence needed above that
PriceGemini 3 Flash Preview: $0.50 per 1M input tokens, $3.00 per 1M output tokens; Batch/Flex $0.25 input and $1.50 outputSelf-hosted: free within licence terms, you pay for compute. Datalab API: $1, $3 or $7.50 per 1,000 pages depending on mode
Output formatsMarkdown, JSON, HTML, plain text, shaped by your promptMarkdown, HTML, JSON and bounding boxes
BenchmarkGemini 3 Flash scores 92.62 on OmniDocBench v1.6 (independent)Chandra 2 scores 85.8 on olmOCR-bench (published by the authors). The two scores come from different benchmarks and cannot be compared directly.
Cost predictabilityToken-based, so cost per page depends on page density and output length. Several newer Flash models rise from $0.75/$3.75 to $1.50/$7.50 per 1M tokens on January 1, 2027.Per-page pricing on the hosted API; self-hosted cost is your compute
Known failureNo fixed schema or confidence scores, since output is prompt-drivenToken-loop hallucinations reported on 5-6 of ~60 Russian/Kazakh scans; 43.2% on Urdu and 51.1% on Sanskrit; Old Scans only 51.1 on olmOCR-bench

Sources

What each one is

Gemini Flash

General vision model with strong handwriting and layout understanding at a low price.

Gemini Flash details

Chandra

Datalab's OCR model that keeps layout and handles forms, tables and handwriting.

Chandra 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": "google/gemini-flash","fallbacks": ["datalab/chandra"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

Gemini Flash alternativesChandra alternativesBest OCR models in 2026