Glimpse
Multimodal model that rechecks images and video mid-inference for physical-world understanding.
Updated 2026-10-19
Yes, within limits. Glimpse runs a free tier with paid plans above it.
Listed pricing: Free tier + $0.60/M input, $3/M output.
Overview
Glimpse is Elorian's multimodal model for reasoning over images and video. It is aimed at researchers and engineers who need answers about physical scenes—what the frames show and what is physically plausible—rather than a general chatbot with vision attached.
The specific technical claim is mid-inference visual re-grounding: the model returns to the image or video evidence while it is still reasoning, instead of treating one upfront encoding as the full visual record. Elorian launched it on October 16, 2026; the launch post and coverage, including a widely shared thread, positioned its benchmarks against GPT-6 Astra. Those comparisons come from launch materials, not from a broad set of independent replications.
I'd use it when the job is an API call that has to look again at visual evidence. I'd skip it when the job is a literature workspace (Elicit, NotebookLM) or a mature general model with a finished app. Treat the Astra comparisons as launch claims until outside evals republish them.
Is Glimpse free?
Yes, within limits. Glimpse runs a free tier with paid plans above it.
What the free tier covers: Free tier available; published caps and rate limits were not included in the launch pricing summary.
Listed pricing: Free tier + $0.60/M input, $3/M output.
Pricing on this page has not been re-verified. The entry was last edited on , and no separate pricing check has been run since. Treat the figures as a record of what was published then and confirm on the official site.
What the free tier leaves out
Read straight off the plan list below. Vendors move features between tiers, so check the current split before you pay.
- API $0.60/M input, $3/M output Usage-based inference. Input is $0.60 per million tokens; output is $3 per million tokens.
Glimpse pricing
| Plan | Price | What's included |
|---|---|---|
| Free | Free | A free tier is offered. Quota, rate limits, and which modalities it includes are not specified in the pricing summary used here. |
| API | $0.60/M input, $3/M output | Usage-based inference. Input is $0.60 per million tokens; output is $3 per million tokens. |
A free tier is offered. Quota, rate limits, and which modalities it includes are not specified in the pricing summary used here.
Usage-based inference. Input is $0.60 per million tokens; output is $3 per million tokens.
Pricing on this page has not been re-verified. The entry was last edited on , and no separate pricing check has been run since. Treat the figures as a record of what was published then and confirm on the official site.
Is Glimpse worth it?
You can test that on the free tier before paying anything. The recorded trade-offs are listed below, and any one of them can settle the question on its own.
The 8.7/10 AI Score is an editorial read of published capability, price and shipping pace. Nobody here has hands-on hours with Glimpse. How we verify.
Worth it if
The strengths recorded against this entry.
- Revisits image and video evidence during inference rather than locking a single upfront encoding
- Covers both still images and video for physical-scene questions
- Published API rates: $0.60 per million input tokens and $3 per million output tokens, plus a free tier
- Scoped as a visual reasoner, so it is easier to judge against general multimodal chat models
Not worth it if
Any one of these blocks your use case.
- Benchmark comparisons with GPT-6 Astra are from the October 16, 2026 launch; independent replication is not established here
- Free-tier quotas, context length, and video-length limits are not stated in the pricing summary
- Output at $3 per million tokens will dominate cost if reasoning traces run long
- No documented note-taking or literature-review workspace; this is a model and API, not a research app
Key features
Visual re-grounding
During inference, Glimpse revisits the image or video instead of answering from a single initial encoding. That is the mechanism Elorian cites for physical-world questions where a one-pass look drops detail.
Image and video
The model takes both still images and video as evidence. It is a reasoning model over visual input, not a generator of images or clips.
Token-priced API
Access is metered at $0.60 per million input tokens and $3 per million output tokens, with a free tier. That suits programmatic use; a separate end-user research app is not part of the published offer.
How it compares
| Tool | Best for | Pricing | Score |
|---|---|---|---|
| Glimpse | — | Free tier + $0.60/M input, $3/M output | 8.7/10 |
| Perplexity AI | Knowledge workers and researchers who want cited, synthesized answers instead of a list of links to click through. | Freemium | 9.4/10 |
| NotebookLM | people who want cited answers and Studio outputs from a specific set of sources. | Free tier; Google AI Pro $19.99/mo | 9.1/10 |
| Inkling-Small | — | Free — open weights download; Tinker Playground access | 8.8/10 |
Compare head-to-head
Comparison explorer Put Glimpse up against any three tools Opens with Glimpse already loaded. Add up to three more from the full index and read pricing, features, pros and cons in one table. →Related reading
DeepSeek V4.1-Flash-1017 Hits Web, App, and API
V4.1-Flash-1017 is live on DeepSeek's site, app, and API, with shorter outputs and a Harness floor of v0.2.1-rc.1.
Anthropic's Public Benefit Fund and the Share Ratchet
Anthropic's fund starts at 5% of shares and can add 2% per 50% value rise. With founder pledges and the employee match, public-bound equity is about 20%.
Run Parallel Agents on GPT-6 Astra Pro
OpenAI's post says multiple agents can split one task in parallel, with progress you follow and finished work you open.
Ready to try Glimpse?
Head to the official site to start with Glimpse — pricing and plans are listed above.
Visit Glimpse