OCR Router
Api

Supported Models

Available Vision LLM models and pricing

Supported Models

DocsRouter supports a curated list of high-performance Vision LLMs. You can list standard available models via the API, but additional models supported by OpenRouter may also serve as valid destination targets.

List Models Endpoint

URL: GET https://api.docsrouter.com/v1/ocr/models

Returns a list of currently active models with their capabilities and pricing.

{
  "object": "list",
  "data": [
    {
      "id": "google/gemini-2.0-flash-001",
      "name": "Gemini 2.0 Flash",
      "cost_per_1m_tokens": 0.10,
      "recommended": true
    },
    ...
  ]
}

Model Reference

IDProviderCost / 1M TokensBest For
google/gemini-2.0-flash-001Google$0.10Default. Best balance of speed, cost, and accuracy for general OCR.
google/gemini-2.5-flashGoogle$0.15Slightly higher accuracy for documents with complex layouts.
openai/gpt-4o-miniOpenAI$0.15Good alternative for simple documents if you prefer OpenAI's reasoning.
anthropic/claude-3.5-sonnetAnthropic$3.00Excellent at handwriting and very dense technical documents/tables.
openai/gpt-4oOpenAI$5.00High-end reasoning, best for extracting "insight" along with text.
google/gemini-2.5-pro-previewGoogle$2.50Google's most capable model.

Selecting a Model

Use the model parameter in your /v1/ocr request:

{
  "model": "anthropic/claude-3.5-sonnet",
  ...
}

If no model is specified, google/gemini-2.0-flash-001 is used.

On this page