# November 2025

> * We improved the accessibility of the **Download advanced metrics** action in the **Measure** tab to support full keyboard navigation and align with WCAG 2.0 Level A requirements.
* **Erratum - added December 4, 2025:** The model used for Document Understanding generative capabilities (including classification, extraction and validation) in the Japan region has been upgraded from `gpt-4o-2024-08-06` to `gpt-4o-2024-11-20`.

## General Release Notes - Document Understanding<sup>TM</sup>

### November 27, 2025

#### Improvements

* We improved the accessibility of the **Download advanced metrics** action in the **Measure** tab to support full keyboard navigation and align with WCAG 2.0 Level A requirements.
* **Erratum - added December 4, 2025:** The model used for Document Understanding generative capabilities (including classification, extraction and validation) in the Japan region has been upgraded from `gpt-4o-2024-08-06` to `gpt-4o-2024-11-20`.

### November 13, 2025

#### Improvements

We have improved how UiPath Document OCR handles high-resolution images. The updated model delivers more consistent results across different image resolutions and improves overall document recognition accuracy.