LightOn Releases LightOnOCR-3 for End-to-End Document Understanding

France's LightOn has released LightOnOCR-3, an end-to-end document understanding model that goes beyond transcription to handle layout and chart parsing, topping several benchmarks. Its predecessor was downloaded over 4.5 million times on Hugging Face, and tests show it can read 16th-century manuscripts and Ancient Greek without task-specific training.

2026-11-02 ~ 2026-11-02 · 2 related posts

• LightOnOCR-3 Launches: Tops FRBench, Adds Layout Structure and Chart Extraction — IgorCarron · 2026-11-02
• LightOnOCR-3 Reads 16th-Century Manuscripts and Ancient Greek Without Training — IgorCarron · 2026-11-02