GLiNER 2.5 Makes Text Extraction Faster and Smarter

The new GLiNER 2.5 model changes how it finds information. Instead of checking every possible word combination, it uses boundary prediction to work much faster. This change means that longer phrases no longer slow down your computer.

You can now use this tool for more complex tasks like connecting entities and identifying specific labels. It supports large documents with up to 4,000 words. You can download three different sizes that are all small enough to run on a standard home computer.

This release offers great results even without extra training. If you need a practical way to pull data from text, this is a strong new option. You can find the code and models available now under an open source license.