Elorian Releases Glimpse AI Model That Reanalyzes Images Mid-Reasoning
Elorian, a Palo Alto startup, released its first AI model, Glimpse, as a research preview with API access and a waitlist-only chat interface. Glimpse uses a "visual scratch space" to crop image regions, pull video frames, or run code mid-reasoning. Elorian's own evaluations claim it outperforms GPT-6 Astra on scientific chart reasoning and short-form video captioning. It costs $0.60 per million input tokens and $3 per million output tokens, targeting robotics, scientific labs, and media companies.
IllustrationEditorial responsibility
- No named human review is recorded for this page.
- Reports are grouped by semantic similarity and deterministic rules. Language models may assist titles, summaries, translation and cross-source analysis; the page reads the event directly, while its address stays stable when the title changes.
- Summary covers the current reports
Cross-source coverage
Reporting timeline
Elorian Releases Glimpse AI Model That Reanalyzes Images and Video Mid-Thought
Elorian has released Glimpse, its first AI model, which can revisit and reanalyze parts of an image or video while reasoning. According to the source, Glimpse can return to a specific time in a video, zoom in on a detail, or replay half a second at a higher frame rate before answering. On Elorian's own tests, Glimpse scored better than GPT-6 Astra on scientific chart reading and short-form video captioning. For planning a simulation of a robot arm placing cream cheese in a bowl, Glimpse cost $0.035 versus $0.249 for GPT-6 Astra. In a demo, Glimpse solved 5 mazes of increasing difficulty, while Gemini 3.8 Flash solved 1/5.
Read sourceElorian Releases First AI Model Glimpse for Visual Re-Examination
Elorian has released its first AI model, Glimpse, described as a visual thinking model that can revisit image or video details mid-reasoning. According to the source, Glimpse can return to a specific video timestamp, zoom into details, or replay half a second of video at a higher frame rate before answering. The announcement positions the model around iterative visual inspection during inference rather than a single-pass response. No performance benchmarks, availability details, pricing, or independent evaluations were provided in the source item.
Read sourceElorianAI's Glimpse Model Compares Video Moments Using Visual Scratch Space
ElorianAI has introduced Glimpse, a model capability that analyzes how a scene changes between points in time in a video and revisits moments for closer inspection when changes are small or fast. Elorian describes this as "visual thinking," meaning the model reasons over what it observes while grounded in the actual frames. During reasoning, Glimpse can return to the source video, crop into a specific area, extract particular frames, or run code to measure and compare what it sees. Elorian refers to this mechanism as a "visual scratch space." The post frames the development as part of a broader trend in which models are becoming significantly better at comparing moments to understand video content.
Read sourceShow 2 older updatesHide older updates
Elorian launches Glimpse, an AI model built to reason with images
Elorian, a Palo Alto AI startup that raised a $55 million seed round at a $300 million valuation before launching a product, released its first model, Glimpse, on Friday as a research preview with API access and a waitlist-only chat interface. The model is built around a mechanism Elorian calls a "visual scratch space," allowing it to crop into image regions, pull specific video frames, or run code mid-reasoning rather than converting an image to tokens once and reasoning in text. CEO Andrew Dai, a former Google DeepMind researcher, cited AI's failure to interpret board game iconography or text-free IKEA diagrams as evidence models still cannot truly see. According to Elorian's evaluations, Glimpse outperforms GPT-6 Astra on scientific chart reasoning and short-form video captioning, with competitive results elsewhere. It costs $0.60 per million input tokens and $3 per million output tokens. Initial targets are robotics firms, scientific labs, and media companies; Elorian aims to build the world's best visual reasoning model next year.
Elorian launches Glimpse, an AI model that reasons with images
Elorian, a Palo Alto AI startup that raised a $55 million seed round at a $300 million valuation before launching a product, released its first model, Glimpse, on Friday as a research preview with API access and a chat interface. The model is built around a mechanism Elorian calls a "visual scratch space," allowing it to crop into image regions, pull specific video frames, or run code mid-reasoning to measure and compare what it sees. According to Elorian's evaluations, Glimpse outperforms even GPT-6 Astra on scientific chart reasoning and short-form video captioning, with competitive results on other visual reasoning tasks. The company says a small team trained Glimpse on a few hundred GPUs, and the model is substantially smaller than the frontier systems it is benchmarked against. Glimpse costs $0.60 per million input tokens and $3 per million output tokens. Initial targets are robotics firms, scientific labs, media, and creative companies, with longer-term goals including direct robot control.