Skip to content
TrustList
All launches
AI model launch

Glimpse

Elorian's visual thinking model, a research preview released 16 October 2026 that reasons over images and video with tools and subagents, at $0.60 in and $3.00 out per million tokens.

Launched 2026-10-16AI model

About this launch

Glimpse is a multimodal model from Elorian, released as a research preview on 16 October 2026 and available the same day through Elorian's API and its chat interface. Elorian calls it a visual thinking model: one that reasons about the physical world from what it sees, rather than turning an image into text and reasoning over the text.

## What it does

Glimpse follows objects in motion and across changing viewpoints, works out how parts fit together, and connects an observation to what is likely to happen next. Elorian says it can revisit and refine internal visual representations while it thinks, so it can zoom into a region, watch a short stretch of video at a denser frame rate, or track one feature over time before answering, and the chat interface shows that visual scratch space. As an agent, it is trained end to end to use multimodal tools that write code, transform images and watch videos, and it can start and coordinate subagents that exchange text, visual artefacts and internal representations.

## Who it is for

Elorian points to robotics, scientific work and the creative industries, with examples ranging from analysing laboratory charts and tagging short social videos to identifying completed actions in robot footage. Its announcement compares Glimpse with larger frontier models on manipulation-event understanding, embodied spatial reasoning, chart reasoning, video captioning, temporal reasoning and counting, claiming it leads on four of those tests and matches on embodied reasoning; those are Elorian's own results under its own harness, and none is stored as a score here. Elorian says the model was built by a small team on a few hundred GPUs and post-trained on an open-weights base model, and that token efficiency is its next priority because the model reads inputs closely and spends tokens doing so.

## Key facts

**Maker:** Elorian **Released:** 16 October 2026 (research preview) **Inputs:** images, video and text **Access:** Elorian API and chat UI **Price:** $0.60 per million input tokens (uncached), $3.00 per million output tokens **Weights:** not published; post-trained on an open-weights base model

Not yet independently verified: Every detail here comes from Elorian's own announcement of 16 October 2026; we have not yet read an independent report. The benchmark comparisons are Elorian's own measurements, so no score is stored. We will update this when it can be confirmed, and remove this note.

Discussion

Questions & answers about Glimpse.

No questions yet. Be the first to ask about Glimpse.