LightOn launches LightOnOCR-3, topping open-weight OCR benchmarks over Mistral OCR 4.1
IgorCarron · x · 2026-10-09
- French company LightOn released LightOnOCR-3, a document-intelligence model family that in one pass handles text/handwriting recognition, image description, chart data extraction, and layout preservation.
- Benchmarks: leads OlmOCR-Bench ahead of Chandra-OCR-2, Mistral OCR 4.1 and dots.mocr; ranks first among open-weight models on ParseBench.
- Ships in compact 0.8B and 4B sizes with up to 2x document processing throughput.
- Positioned for Europe's large backlog of document digitalization; a strong OCR model effectively enables ctrl+F over enterprise data.
More from Multimodal
- Iris-3B: a 3B pixel-space text-to-image model with no VAE, fully open under Apache 2.0 — zhenjun_zhao · 2026-10-09
- Voyager launches: an open harness that plugs frontier models into creative tools — testingcatalog · 2026-10-09
- Voyager launches as an open harness driving Opus, Astra and DeepSeek across AE, Blender and more — HeyAmit_ · 2026-10-09
- YuE2 LoRA Training for New Genres Keeps Failing, Redditor Asks for Help — EuphoricTrainer311 · 2026-10-09
- Long Shot Studio: an open-source ComfyUI UI for multi-shot long video with resumable rendering — R34vspec · 2026-10-09
- 30-second cinematic creature short made with Seedance 2.5 draft mode upscaled to 1080p — LudovicCreator · 2026-10-09