AI News · Models ·
Elorian releases Glimpse, a visual reasoning model priced at $0.60 per million tokens
Palo Alto startup Elorian released its first model, Glimpse, as a research preview with API access. The model can crop images, pull video frames or run code while it reasons. Elorian previously raised a $55 million seed round at a $300 million valuation.
Key points
- Glimpse launched as a research preview with API access; chat is waitlist only.
- Pricing is $0.60 per million input tokens and $3 per million output tokens.
- Elorian raised a $55 million seed round at a $300 million valuation.
What happened: Elorian, a Palo Alto AI startup, released its first model, Glimpse, as a research preview with API access, TechCrunch reported. A chat interface is available by waitlist only. The company raised a $55 million seed round at a $300 million valuation before launching a product, co-led by Striker Venture Partners, Menlo Ventures and Altimeter Capital, with participation from Nvidia and Google chief scientist Jeff Dean.
The details: Unlike models that look at an image once and then reason in text, Glimpse can return to the source during reasoning. It can crop regions, pull specific video frames or run code to measure and compare what it sees, which Elorian calls a "visual scratch space." Pricing is $0.60 per million input tokens and $3 per million output tokens. Elorian said a small team trained the model on a few hundred GPUs and that it is substantially smaller than the frontier systems it is compared against.
Background: In its own evaluations, Elorian said Glimpse outperforms GPT-6 Astra on scientific chart reasoning and short-form video captioning, with competitive results on other visual tasks. CEO Andrew Dai, a former Google DeepMind researcher, said AI models struggle with visual language such as board game symbols and IKEA diagrams.
Who it affects: Initial customers are companies with large video libraries, including robotics firms, scientific labs, and media and creative companies. Longer term, Elorian wants to move from seeing to acting, including controlling robots directly.
Our take
Teams that analyze charts, video or diagrams get a new mid-priced API option, though its benchmark claims come from Elorian's own tests.