Glimpse
Elorian's first model: a visual reasoning model that crops, pulls video frames, and runs code on the source image while it thinks.
Overview
Elorian released Glimpse on October 16, 2026 as a research preview with API access and a chat interface. Rather than converting an image to tokens once, Glimpse revisits the source image during reasoning, using a 'visual scratch space' to crop regions, extract video frames, or run code to measure and compare what it sees. Elorian says it beats GPT-6 Astra on scientific chart reasoning and short-form video captioning, but these results are company-reported and not independently verified. The startup raised a $55M seed at a $300M valuation, co-led by Striker Venture Partners, Menlo Ventures, and Altimeter Capital, with participation from NVIDIA and Jeff Dean.
This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.
Key features
- Revisits the original image or video during reasoning instead of a one-time tokenization
- Visual scratch space: crop regions, extract video frames, run code to measure or compare
- Research preview with API access and a chat interface (waitlist reported for chat)
- Company-reported wins over GPT-6 Astra on scientific chart reasoning and short-form video captioning
- Trained by a small team on a few hundred GPUs, per the company
Use cases
- Scientific chart and figure analysis
- Video-heavy workflows in robotics, media, and creative teams
- Visual evidence checking where measurement or zoom matters
Pricing
API: $0.60 per million input tokens and $3 per million output tokens (research preview).
Pricing reflects the source record and may change. Confirm current terms with the provider.
Dates and sources
- Released
- 16 Oct 2026
- Record updated
- 18 Oct 2026
- Record last checked
- 18 Oct 2026