Instance is live · custom pricing · last verified 4 Nov 2026 · viability monitor
Instance
A Y Combinator-backed research effort working toward a standard benchmark for world models and video generation
There is no product here yet — a landing page, two email addresses, and a YC badge. We would not plan any evaluation work around Instance today, and a roadmap that assumes a working benchmark from them is a bet on a promise, not a purchase. If you need video-model evaluation this quarter, published academic benchmarks and the scoring harnesses shipped alongside them are what you actually run. Track Instance as a name; check back once a methodology paper or metric definitions exist.
Verdict generatedVerified 10d ago · liveness 59/100 · cite: rightaichoice.com/tools/instance
- AI researchers tracking early-stage benchmarking efforts
- World-model builders watching for a shared evaluation framework
- Evaluation teams keeping a watching brief on new eval methodology
- Teams that need a working video evaluation framework today — nothing is released
- Anyone looking for a video generation tool — Instance is not a generator
- Non-technical users wanting hosted, plug-and-play model scoring
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Instance if you need reproducible video-model evaluation scores this quarter — the public site currently publishes only "We're working on world models," a YC badge and two contact addresses, with no methodology or metric definitions to audit.
Instance's pricing fits teams whose volume aligns with the published tiers. Compare against the alternatives listed below for stage-specific value.
In short
Instance — A Y Combinator-backed research effort working toward a standard benchmark for world models and video generation. Best for AI researchers tracking early-stage benchmarking efforts, World-model builders watching for a shared evaluation framework, Evaluation teams keeping a watching brief on new eval methodology. Pricing is on request (custom quotes).
What is Instance?
Instance — A Y Combinator-backed research effort working toward a standard benchmark for world models and video generation. Best for AI researchers tracking early-stage benchmarking efforts, World-model builders watching for a shared evaluation framework, Evaluation teams keeping a watching brief on new eval methodology. Pricing is on request (custom quotes).
Instance is a Y Combinator-backed research initiative working toward a standardized benchmark for video generation models and world models. The problem it names is real for anyone who has tried to compare generative video systems: evaluations today lean on cherry-picked demo reels and metrics that different labs define differently, so numbers don't travel between teams. The stated ambition is a repeatable, comparative framework — run several models through the same pipeline and read results side by side across things like temporal coherence, visual fidelity, and semantic alignment with the input prompt, with a pipeline designed to scale across video lengths. Who that is for is narrow: research teams, world-model builders, and evaluation groups inside labs or studios who need comparative numbers rather than another generator to play with. What exists today is much thinner than the ambition. As of this refresh the live site at tryinstance.app is a single static page that repeats one line — "We're working on world models" — alongside a Y Combinator badge, two contact addresses (claire@tryinstance.app, lucy@tryinstance.app), and LinkedIn and Twitter links. No methodology paper, no metric definitions, no public leaderboard, and no usage surface are published there. Everything described in the paragraph above is planned, not shipped.
Key features
- Y Combinator-backed research initiative (per site badge)
- Stated focus on world models
- Planned standardized evaluation framework for video generation models
- Planned temporal coherence metrics for generated video
- Planned visual fidelity scoring
- Planned semantic alignment scoring between prompt and output video
- Planned side-by-side comparative scoring of multiple models
- Planned reproducible evaluation methodology
- Planned scalable evaluation pipeline for diverse video lengths
- Public contact addresses claire@tryinstance.app and lucy@tryinstance.app
- LinkedIn and Twitter presence linked from the site
Behind the verdict
Instance sits in an unusual category for a review: it is a company with a thesis, not a company with a product. The thesis is defensible. Comparing video generation and world models is genuinely hard, because labs publish highlight reels rather than comparable numbers, and every team's internal eval uses its own prompts, its own raters and its own definition of "temporal coherence." A shared harness that runs multiple models through one pipeline and reports temporal coherence, visual fidelity and semantic alignment side by side would be useful to anyone who has to make a build-versus-buy call on a video model. What the public site supports is much narrower than that thesis. The scrape of tryinstance.app returns one repeated sentence — "We're working on world models" — plus a Y Combinator badge, the two contact addresses claire@tryinstance.app and lucy@tryinstance.app, and LinkedIn and Twitter links. There is no methodology page, no metric definition, no leaderboard, no dataset card and no way to run anything. We could not reach documentation or developer pages this run, so we make no claim about what sits behind them; equally, nothing in the public page lets us verify a single capability. That gap matters for procurement. An evaluation framework is only useful if you can read its definitions and reproduce its scores, because a benchmark you can't audit is just someone else's opinion with a number attached. Until Instance publishes the metric definitions and the pipeline, its numbers can't be checked, challenged or cited — which is precisely the problem it says it wants to solve. Where it fits: as a name to watch if your team builds or evaluates world models and you want to know whether a shared benchmark is emerging. A researcher tracking the space can follow the LinkedIn and Twitter accounts for a methodology announcement. Where it doesn't: any team with a decision to make this quarter, any team that needs reproducible scores now, and any team that needs a priced, contractual vendor relationship. If you are choosing a video model today, run the published academic benchmarks and the scoring harnesses that ship with them, and build your own small prompt set on top — that will outperform waiting on anyone's roadmap. The honest summary: a credible team and a real problem, with nothing public yet to evaluate them against. Revisit when there is a paper.
Limitations
- As of this refresh the public site is a minimal single page: the line "We're working on world models," a Y Combinator badge, two contact addresses, and LinkedIn and Twitter links.
- No methodology paper, no metric definitions, no leaderboard, and no usage surface are published there, so the evaluation framework and its metrics remain unverifiable from the evidence.
- The source fetch did not reach any documentation or developer pages, so we make no claim about what does or does not exist behind them.
- Treat every capability named above as planned rather than shipped.
as of 2026-10-30
Researching Instance? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is Instance still active?
Instance is live · custom pricing · last verified 4 Nov 2026 · viability monitor.
Viability score
How well maintained and how widely used is Instance? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: November 2026
How we score →What do users say about Instance?
What people actually say about Instance — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
ResearchedDiscussed by users across 6 sources (Hacker News, YouTube, Product Hunt, Bluesky, Stack Overflow, Lemmy) · researched Nov 13, 2026.
Weighted by the 119 posts each of 6 sources contributed.
- +Targets a real, well-named problem: video-model evals rely on incompatible, lab-specific metrics
- +YC backing and badge give early credibility that may attract research collaborators
- +Proposed multi-axis scoring (temporal coherence, fidelity, semantic alignment) aligns with what research teams actually need
- +Planned side-by-side comparative scoring across several models addresses a real gap in published evals
- +Ambition to scale across diverse video lengths is a meaningful differentiator versus fixed-clip benchmarks
- −No methodology paper, metric definitions, leaderboard, or runnable code is published — everything is planned
- −Zero genuine community discussion; almost all scraped posts are coincidental keyword matches to other products
- −Generic name collides with a Product Hunt app builder and countless technical uses of 'instance'
- −No evidence of responsiveness, uptime, or maintenance because no public issues or user base exist
- −Promised reproducibility and comparative numbers are unverifiable without artifacts
- • No published pricing means any engagement requires direct negotiation with unknown terms
- • Unquantified integration and evaluation-setup cost for teams wanting to run comparisons
How we verified this
Every listed fact about Instance — pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 4 Nov 2026. Analysis is drafted with AI models and checked against our editorial standards before publishing; see how we verify.
Verification history
We have re-verified Instance 16 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed setup time, pricing fit, who should skip it, summary
- — changed who it suits, description, our verdict, features, limitations, who should skip it, pricing fit, setup time, summary, use cases
- — changed limitations
- — changed description, summary, who it suits, who should skip it, our verdict, features
- — changed SSL certificate
- — changed domain expiry
Showing the 6 most recent of 16 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does Instance cost in 2026?
Instance: Pricing is on request (custom quotes). Pricing last verified 4 Nov 2026.
Where the pricing makes sense
The company stage and team size where Instance's pricing actually pencils out — and where peers do it cheaper.
Instance's pricing fits teams whose volume aligns with the published tiers. Compare against the alternatives listed below for stage-specific value.
Has Instance's pricing changed?
No plan price for Instance has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick Instance?
Pick Instance if you are AI researchers tracking early-stage benchmarking efforts, World-model builders watching for a shared evaluation framework, Evaluation teams keeping a watching brief on new eval methodology.
Look elsewhere if you are Teams that need a working video evaluation framework today — nothing is released, Anyone looking for a video generation tool — Instance is not a generator, Non-technical users wanting hosted, plug-and-play model scoring, Procurement teams requiring a priced, contractual product.
Skip it if Skip Instance if you need reproducible video-model evaluation scores this quarter — the public site currently publishes only "We're working on world models," a YC badge and two contact addresses, with no methodology or metric definitions to audit.
GeneratedReal-world workflow fit
Concrete scenarios for the personas Instance actually fits — and what changes day-one when you adopt it.
You want to know whether a shared, auditable benchmark for video generation is emerging before you commit your team's internal eval to someone else's metric definitions. You open tryinstance.app, find the single "We're working on world models" page, note the Y Combinator badge, and follow the linked LinkedIn and Twitter accounts.
Outcome: You get a name to track and a way to hear about a methodology release, at the cost of about two minutes — but no numbers to cite and nothing to run.
You need to compare candidate video models on temporal coherence and prompt alignment for a Q4 production decision and you check whether Instance's framework can be used as the common yardstick.
Outcome: You find no published metric definitions or pipeline, so you fall back to published academic benchmarks plus a small internal prompt set — and keep Instance on a watchlist rather than in the decision.
Use cases
- Tracking an early-stage YC-backed effort to standardize video-model evaluation
- Following plans for a shared benchmark for world models (nothing runnable yet)
Setup time & first value
How long it actually takes to get something useful out of Instance — broken out by persona, not the marketing-page minute.
There is nothing to set up. Reading the entire public page takes under two minutes: one repeated sentence, a Y Combinator badge, two contact addresses (claire@ and lucy@tryinstance.app) and LinkedIn/Twitter links. Following the social accounts to catch a future methodology release takes seconds.
Resources & guides
Tutorials & Learning
YouTube returned 6 videos for “Instance”, and we withheld 6: 6 could not be judged, because “Instance” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Instance.
Official links
What are the best Instance alternatives?
The closest alternatives to Instance we list are Kapwing, CapCut and Palmier Pro. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| Kapwing | Freemium | Pricing not published | — |
| CapCut | Freemium | Pricing not published | — |
| Palmier Pro | Freemium | Pricing not published | — |
Alternatives to Instance
View allKapwing
Browser-based AI video editing with auto-subtitles, dubbing, and prompt-to-video generation
CapCut
ByteDance's browser, desktop and mobile video editor with templates, auto-captions and AI clip generation in one timeline.
Palmier Pro
Palmier Pro is a macOS AI video editor where Claude, Codex, or Cursor cut your timeline over MCP and generation lands in the edit.
Tools that pair well with Instance
Common stack mates teams adopt alongside Instance, with the specific reason each pairing earns its keep.
Kapwing
Browser-based AI video editing with auto-subtitles, dubbing, and prompt-to-video generation
CapCut
ByteDance's browser, desktop and mobile video editor with templates, auto-captions and AI clip generation in one timeline.
Palmier Pro
Palmier Pro is a macOS AI video editor where Claude, Codex, or Cursor cut your timeline over MCP and generation lands in the edit.
Instance vs its closest rival
Instance's closest head-to-head is with Splice.
Our Instance vs Splice verdict written Jul 3, 2026 is more than 90 days old, so it is being re-reviewed against current data and is not shown.
Frequently asked questions
Categories
Best-of guides
Topics
For makers of Instance
- Get a Verified Listing for Instance ($39): a written decision within 1 working day.
- Free listing for another tool: Free reviews are decided in order, within 3 weeks of the review starting.
Used Instance? Help shape our editorial sentiment research.