- Overview
- Model building
- Model validation
- Model deployment
- Consuming models
- API
- Frequently asked questions
- Access and availability
- Data and security
- Migration
- Design-time
- Runtime
- Document processing limits
- Maintenance and monitoring
In practice, IXP can reliably extract up to approximately 1,200 field values per document.
This means that if a document requires more than 1,200 field values to be extracted, the model may reach its token limit. This can happen even if the document is well within the page limits listed in this section. Research and development work is ongoing to support higher field value limits in future releases.
What happens when limits are exceeded?
If token limits are exceeded, predictions may be incomplete or empty regardless of the number of pages. For example, an 80-page document with 1,400 extracted field values may trigger a token-limit warning in the user interface. At runtime, when the token limit is exceeded, the API response will show this by returning occurrence and extraction confidence values of 0.
If page limits are exceeded, only the pages within the specified limits below are processed. Both of these behaviors apply to both Gemini and GPT-4o.
Runtime limits include:
- Number of pages per document:
- For Gemini, up to 500 pages.
- For GPT-4o, 50 pages.
- Token limits - Data-heavy documents with more than 200 extractions, such as lengthy tables, will likely reach the token limit if they have a typical number of fields per extraction. Gemini has a higher output token limit by default, so it can return a larger number of field values in a single call.