Elorian, an artificial intelligence startup based in Palo Alto, released its first model on Friday. The model is called Glimpse, and it is launching as a research preview.
Users can access Glimpse through an API and a chat interface. TechCrunch first reported the release.
The company raised a $55 million seed round at a $300 million valuation before it had a product. Its main idea is that today’s AI models are good with language and code but still struggle to truly see.
How Glimpse Works
Most multimodal models look at an image once. They turn it into tokens and then reason about it in text, a bit like describing a photo from memory.
Glimpse takes a different path. While it reasons, it can go back to the image, crop into a region, pull specific frames from a video, or run code to measure what it sees.
Elorian calls this feature a “visual scratch space.” It works much like a person zooming in on a phone photo to check a detail.
Elorian CEO Andrew Dai, a former Google DeepMind researcher, said the idea grew from a game night last year. He took a photo of a finished board game and asked a leading AI model who was winning, and it could not answer.
“To an adult, it’s quite easy to understand what’s going on with minimal text, but these models aren’t able to intuit what the symbols mean,” Dai told TechCrunch.
He also pointed to furniture instructions. “Humans can look at Ikea diagrams, with no text, and assemble Ikea furniture, right? Why can’t AI models?” he said.
The approach is not entirely new. OpenAI added crop-and-zoom during reasoning to its o-series models in 2025.
Elorian says the difference is that it built its whole model around this ability. The company says a small team trained Glimpse on a few hundred GPUs, and the model is much smaller than the frontier systems it was tested against.
Pricing, Performance, and Backers
Glimpse costs $0.60 per million input tokens and $3 per million output tokens. That is cheaper than mid-level models like Sonnet and Sol, but more expensive than lighter models like Haiku and Luna.
According to Elorian’s own tests, Glimpse outperforms GPT-6 Astra on scientific chart reasoning and short-form video captioning. The company says it posts competitive results on other visual reasoning tasks.
Elorian came out of stealth six months ago. Its funding round was co-led by Striker Venture Partners, Menlo Ventures, and Altimeter Capital, with Nvidia and Google chief scientist Jeff Dean also taking part.
The company is first targeting businesses with large amounts of video. These include robotics firms, scientific labs, media outlets, and creative companies.
Over time, Elorian plans to move from seeing to doing. That means building a model of a scene and using it to choose actions, including directly controlling robots.
Dai said the company aims next year to build “the world’s best visual reasoning model,” one that leads on every visual reasoning test.
