model fieldA compact 0.9B-parameter vision-language model purpose-built for document parsing, pairing a dynamic-resolution visual encoder with a lightweight ERNIE-4.5 language model. We run version 1.6, which scores 96.33% on OmniDocBench v1.6 (state-of-the-art) and adds seal/stamp recognition, text spotting, and improved Chinese ancient-document and rare-character handling on top of the base model's tables, formulas, and charts across 109 languages. Apache 2.0.
Input Price
₹0.02
per 1K input tokens
Output Price
₹0.02
per 1K output tokens
Context Window
16Ktokens
paddleocr-vlmax_tokensisn't set)The model's own limit may be lower — not published.OpenAI-compatible. Replace nbh_YOUR_API_KEY with a key from your account.
curl https://api.nabh.cloud/v1/models/paddleocr-vl/chat/completions \
-H "Authorization: Bearer nbh_YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "paddleocr-vl",
"messages": [
{
"role": "user",
"content": "Hello!"
}
]
}'Free account to start · your first ₹500 recharge unlocks this model
Prepaid, in INR. Charged per request from your credits.
Per account, across all models. Over the limit, requests get HTTP 429 — back off and retry.
API requests (with your key or from the playground) are passed to the model to generate the response and are not stored by nabh.cloud — no prompts, no outputs. We keep request metadata (time, model, tokens, latency, cost) for billing and your request log.