Elorian releases Glimpse, a visual reasoning model, as a research preview
Article timeline
The articles in this story. Times are ET.
Rohan Paul@rohanpaul_aiXElorian's Glimpse model compares video frames to reason about changesAIElorian releases Glimpse, a visual thinking model that compares how a scene changes across video moments. When the change is small or fast, Glimpse returns to the source, crops regions, extracts frames, or runs code to measure what it sees, which Elorian calls a "visual scratch space." Elorian says Glimpse is competitive with Astra on key public visual reasoning benchmarks.
Santiago@svpinoXGlimpse answers questions about images and videos, including spatial relations and motionAISantiago says a model called Glimpse can answer questions about an image or video, including how objects relate spatially and how an action changes over time. Andrew Dai introduces Glimpse as a visual thinking model with competitive frontier performance against Astra on key public visual reasoning benchmarks.
Jeff Dean@JeffDeanXGlimpse, Elorian's first visual thinking model, targets moving and angle-shifting imagesAIAndrew Dai introduces Glimpse, a visual thinking model from Elorian that he says performs competitively with Astra on key public visual reasoning benchmarks. He says earlier models handled single images well but struggled with anything that moves or changes angle. Jeff Dean praises the work of Dai and the Elorian team.
elvis@omarsar0XGlimpse, a new visual model, scores 8/10 on board-reading test vs 5/10 for Gemini 3.8 FlashAIElorian AI releases Glimpse, its first vision-language model, which its founders say led work at Google on Gemini and at Apple on MM-1. In a 10-image test of game boards from an overhead camera, the post says Glimpse read 8 boards fully correctly against 5 for Gemini 3.8 Flash, with results hand-checked.
- TechCrunch · AINewsElorian releases Glimpse, a visual reasoning model that re-examines images mid-reasoning
AIElorian releases Glimpse, its first model, as a research preview with API access, priced at $0.60 per million input tokens and $3 per million output tokens. Glimpse can crop image regions, pull video frames, and run code during reasoning, which Elorian calls a "visual scratch space." The company says Glimpse outperforms GPT-6 Astra on scientific chart reasoning and short-form video captioning, according to its own evaluations.