Elorian’s first AI model wants to reason with images, not just about them
According to Elorian's evaluations, Glimpse outperforms even GPT-6 Astra on scientific chart reasoning and short-form video captioning, with competitive results on other visual reasoning tasks.
TL;DR
- Elorian, an AI startup, has released its first model called Glimpse.
- Glimpse focuses on deep visual reasoning by continuously referring back to the source image or video.
- It allows actions like cropping, zooming, and running code on visual data during reasoning.
- The model was developed with a focus on cost-efficiency and a smaller footprint compared to other frontier systems.
- Elorian aims to target companies with large amounts of video data and eventually expand to robotics and direct action control.