Elorian, founded by ex-DeepMind researcher Andrew Dai, releases first visual reasoning model Glimpse
Palo Alto startup Elorian launched Glimpse research preview on Oct 16, introducing a 'visual scratch space' mechanism, priced at $0.60 input / $3.00 output per 1M tokens.
On October 16, Palo Alto AI startup Elorian released Glimpse, its first model, as a research preview with API access and a waitlist-only chat interface. Elorian was founded by former Google DeepMind researcher Andrew Dai. The company emerged from stealth earlier this year with a $55 million seed round at a $300 million valuation, backed by Striker Venture Partners, Menlo Ventures, Altimeter Capital, Nvidia, and Google chief scientist Jeff Dean.
Glimpse is built around a visual scratch space mechanism: unlike most multimodal models that look at an image once, convert it to tokens, then reason about it in text, Glimpse keeps going back to the source — cropping into a region, pulling specific frames from a video, or running code to measure and compare what it is looking at.
API pricing is $0.60 input and $3.00 output per 1M tokens, between mid-tier models like Sonnet and Sol and lightweight models like Haiku and Luna. In Elorian's evaluations, Glimpse outperforms GPT-6 Astra on scientific chart reasoning and short-form video captioning, with competitive results on other visual reasoning tasks.
Glimpse was trained by a small team on a few hundred GPUs; the company says the model is substantially smaller than frontier systems it was benchmarked against.