AI Startup Elorian Releases First Model Glimpse With a 'Visual Scratch Space'

On October 16, Palo Alto-based AI startup Elorian released its first model, Glimpse, as a research preview with API access; the chat interface remains waitlist-only. The company came out of stealth around April 2026 and had raised a $55 million seed round at a $300 million valuation before shipping a product.

Glimpse is built around what Elorian calls a "visual scratch space." Unlike most multimodal models, which convert an image into tokens and then reason about it in text, Glimpse repeatedly returns to the source image during inference: it can crop into a region, pull specific frames from a video, or run code to measure and compare what it sees.

CEO Andrew Dai, a former Google DeepMind researcher, said the approach stems from frontier models struggling with visual symbols, such as those in board games. Elorian said a small team trained Glimpse on a few hundred GPUs, making it substantially smaller than the frontier systems it benchmarks against. Pricing is set at $0.60 per million input tokens and $3 per million output tokens.

According to 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. The company is initially targeting businesses with large amounts of video, including robotics firms, scientific labs, media, and creative companies.

Longer term, Elorian wants to move from seeing to doing, building a model of a scene and using it to choose actions, including controlling robots directly. Dai said the company aims to build "the world's best visual reasoning model" by 2027.