Elorian releases Glimpse model for reasoning through images and video
By Automate Basics, written with AI from TechCrunch's original
Elorian has released Glimpse, a research-preview AI model that can inspect parts of an image or video again while working out an answer. API access is available for teams that handle visual material, while its chat interface has a waitlist.
What changed with Glimpse
Elorian has released Glimpse, its first AI model, as a research preview with API access. TechCrunch reports that the model can return to visual material while forming an answer, rather than relying only on an initial reading of an image. Elorian says Glimpse can examine a cropped area, select frames from a video or use code to measure and compare details.
That approach could matter when an answer depends on a small feature in a chart or a brief moment in a clip. In Elorian's own evaluations, Glimpse performed well on scientific chart reasoning and captioning short videos, including against a larger model. Those are company-reported results, not a guarantee that it will interpret every workplace image or video correctly.
Who can use Glimpse, and what does it cost?
Glimpse is available through an API research preview, according to TechCrunch. Its chat interface is waitlist-only, so someone looking for a ready-to-use chat tool cannot assume they can use that interface now. The report does not name any subscription plans or say whether access to the API requires approval.
TechCrunch reports a price of $0.60 per million input tokens and $3 per million output tokens. Those are usage rates, not a quoted monthly bill. Elorian is initially targeting organizations with substantial video collections, including scientific labs, media and creative companies. For a team without someone who can work with an API, the chat waitlist is the relevant availability limit.
How do I try Glimpse at work?
Glimpse is worth evaluating on a visual task where a wrong detail would be easy for a person to spot. A team with API access could start with a scientific chart that has several closely spaced lines, or a short video whose important event appears only briefly. Ask for a description or answer that depends on the visual evidence, then compare it with the original material.
Check whether the answer identifies the right part of the chart or moment in the clip, and whether it adds details that are not visible. Keep a person responsible for the final interpretation. TechCrunch describes Glimpse as a research preview, and the performance figures come from Elorian's evaluations; the report does not establish how reliably it handles a particular team's files.
Frequently asked questions
Can I use Glimpse in a chat interface?
Not without joining a waitlist, according to TechCrunch. Glimpse has launched as a research preview with API access, but its chat interface is waitlist-only. The report does not say when broader chat access will begin or describe a subscription plan for it.
What kinds of visual work is Glimpse meant to handle?
Elorian says Glimpse can revisit parts of an image, select video frames and use code to compare what it sees while answering. The company reports strong results on scientific chart reasoning and short-video captioning. Those claims are based on Elorian's evaluations, so check its answers against your own material.
How much does Glimpse cost?
TechCrunch reports usage prices of $0.60 per million input tokens and $3 per million output tokens. The source does not provide a monthly subscription price or estimate the cost of a typical chart or video task, so a team's bill would depend on its usage.
For you
Learn to do this with AI, free
Short courses for people who aren't engineers: ChatGPT, Claude, prompting and automation. Every lesson is free, with no card.
Start a free courseFor companies
Make an AI or automation tool?
List it in the directory for $9.99, refunded in full if it isn't approved, or claim its page if it is already here.
Related reading
Veritas expands verification chip to jewelry and glass bottlesVeritas has introduced a chip that can verify products on metal surfaces and through glass, according to TechCrunch. The update gives brands more ways to build authenticity checks into goods such as jewelry, trophies and fragrance bottles.
Meta and Apple work together to run Muse Spark Open on MacsMeta says its new open-weight Muse Spark Open model can run on a pair of Apple M5 Ultra Mac Studio computers with the latest MLX framework. The setup gives workplace teams a local option to assess, though the announcement does not explain its cost.
Bluesky expands Attie to create websites, quizzes and gamesBluesky has expanded its experimental Attie platform beyond custom social feeds. TechCrunch reports that people can now use it to create websites, quizzes, games and other interactive projects connected to the Bluesky network.
Written with AI from TechCrunch's original and published after automatic checks: every figure here appears in the original, and no sentence is copied from it. The picture is AI-generated. The original is the authority.
Get the next one by email
One email, most weeks: what changed in AI for people who use it at work, and what to do about it.
Source: Elorian’s first AI model wants to reason with images, not just about them, TechCrunch, 16 Oct 2026.