# Decart's Oasis 3 and the wager on 'world models'

> On 10 June 2026 Decart launched Oasis 3, a real-time world model for driving simulation, via API.

*A photorealistic driving simulator you can prompt in plain language — and an attempt to do for world models what OpenAI did for language.*

By WireRead Editorial · WireRead
Canonical: https://wireread.com/news/decart-oasis-3-world-model-wager

A *world model* is not a chatbot or an image generator. It is an AI that builds an **interactive environment** you can act inside — one that responds as you move through it, generating the next frame based on your actions rather than retrieving a stored clip. **Decart's Oasis 3**, launched 10 June 2026, is the most commercially pointed version yet: a model that spins up photorealistic driving scenes in real time from a text prompt, served via API. The ambition Decart is selling is not simply the demo — it is the *platform*.

## What Oasis 3 actually does

Feed Oasis 3 a natural-language description and it generates a **multi-camera driving scene** — one front-facing view, two side — rendered photorealistically and explorable in real time. Weather, road surface, traffic density, lighting conditions: all variable, all promptable. The primary use case is **autonomous-vehicle training**, where teams need to rehearse edge-case scenarios at scale — black ice at 3 a.m., a cyclist appearing from behind a parked lorry, a flooded junction — that are rare enough never to be reliably encountered on real roads and too dangerous to stage repeatedly.

Decart prices the API at roughly **$0.02 per second** of simulated time. That is cheap enough that a team can generate thousands of scenario variants overnight; it is not free, which means Decart is trying to build a real business, not a demo. The company says robotics and broader 'physical AI' applications are the next expansion, once the driving domain is proven.

> Decart describes Oasis 3 as laying the foundation for physical AI systems — a world model that can simulate hours of photorealistic driving from natural-language prompts, available via API for autonomous-vehicle teams on day one.
> — [Dataconomy](https://dataconomy.com/2026/06/10/decart-lays-the-foundation-for-physical-ai-systems-with-oasis-3/), 2026-06-10

## The platform bet

The strategic framing CEO **Dean Leitersdorf** is pushing is deliberately reminiscent of the early OpenAI API era: 'the first usable world model that people can actually program on top of'. The claim is not that Oasis 3 is the most impressive world model ever built — it is that it is the first *accessible* one, with pricing, an API, and a focused domain that makes it immediately actionable for a paying customer base.

That is a structurally significant distinction. Language models became a platform because developers could build products on them cheaply, with no ML expertise required. If world models follow the same trajectory — and Decart is making a $300m-backed bet that they will — the value accrues to whoever owns the layer everyone else's physical-AI applications run on. That makes the developer ecosystem more important than the demo reel, and the pricing more important than the benchmark score.

> **Key:** **The throughline is the platform, not the product.** Decart is not trying to be the best driving simulator — it is trying to be the API layer that autonomous-vehicle and robotics teams build their simulation stacks on. Whether that bet pays depends on whether the developer community shows up, and on how fast the real-world fidelity gap closes.

## The caveat that is the whole ballgame

Multiple outlets that stress-tested Oasis 3 flagged **real-world limitations** in the simulations. This matters more here than in almost any other category. A photorealistic world model for entertainment can have visible glitches — that is fine. A world model being used to train a self-driving car that will subsequently operate on public roads cannot. The **sim-to-reality gap** is precisely the failure mode that AV training must guard against: a model that learns from slightly wrong physics, slightly wrong depth, or slightly wrong edge-case behaviour does not fail in the simulator — it fails on a real road at 60 mph.

'Photorealistic' is not 'physically accurate'. Decart has not claimed full accuracy; the caveats in the coverage are honest. What matters is whether the company can close the gap fast enough that the simulations teach *correct* behaviours rather than plausible-looking ones. The ~$300m raise at ~$4bn — with Toyota and Nvidia among the backers — buys runway to get there. Toyota's presence in the cap table is notable: an AV-adjacent manufacturer investing in a world-model startup is a signal of strategic intent, not just financial yield.

The comparison with language models is instructive here too. Early LLM APIs hallucinated constantly — that was tolerable for text generation and fatal for medical diagnoses. The world-model version of that same problem is a simulator with plausible-but-wrong physics. Which domains can tolerate imperfect simulation, and which cannot, will determine how quickly Decart's platform thesis can actually be tested at scale.

| Dimension | Current state | What changes the calculus |
| --- | --- | --- |
| **Fidelity** | Photorealistic, limited real-world accuracy | Physics-accurate sim-to-reality transfer |
| **Domain** | Driving only | Robotics + open-environment expansions |
| **Pricing** | ~$0.02/second | Volume discounts as developer base grows |
| **Ecosystem** | API day one, developer community TBC | Adoption by AV/robotics teams at scale |
| **Competition** | Incumbent in-house AV simulators | Other world-model startups entering the space |

The table above maps what Oasis 3 is *now* against what the thesis requires. The honest read: the right product at the right time, if the fidelity improves.

> Decart's world model can simulate hours of photorealistic driving — but comes with real-world caveats that matter most in safety-critical autonomous-vehicle training.
> — [TechCrunch](https://techcrunch.com/2026/06/10/decarts-new-world-model-can-simulate-hours-of-photorealistic-driving-with-some-caveats/), 2026-06-10

## What to watch next

Two things will determine whether the platform bet lands. First, **developer adoption**: whether AV and robotics teams actually integrate Oasis 3 into their pipelines, or treat it as a demo and stick with bespoke in-house simulators. Second, **fidelity progress**: whether Decart's next releases close the sim-to-reality gap enough that the simulations are safe to use for safety-critical training, not just exploration.

The robotics expansion is worth watching separately. Driving is a well-defined domain — camera angles, road physics, a finite set of actors. General robotics is vastly harder: unstructured environments, contact physics, manipulation. If Oasis 3's architecture generalises cleanly to manipulation tasks, the world-model platform thesis becomes much larger than AV training. If it does not, Decart has built a well-funded specialist tool in a category that is already served by the AV industry's mature in-house simulators. The answer will arrive in the next 12–18 months of developer usage, not in the launch press release.

## Key takeaways

- Oasis 3 generates photorealistic, interactive multi-camera driving scenes in real time from text prompts, available via API at $0.02/second.
- The first target market is autonomous-vehicle teams who need to rehearse rare or dangerous scenarios at scale without waiting to encounter them on real roads.
- CEO Dean Leitersdorf frames it as 'the first usable world model that people can actually program on top of' — a platform bet, not just a driving tool.
- Multiple outlets that tested Oasis 3 flagged real-world limitations; 'photorealistic' and 'perfect' are not the same, and the gap matters most in safety-critical simulation.
- Decart raised ~$300m at ~$4bn weeks before the launch, with Toyota, Adobe, eBay and Nvidia among backers — runway to close the sim-to-reality gap if the developer ecosystem materialises.

## FAQ

### What is a world model, and how is Oasis 3 different from a video game?
A world model is an AI that generates an interactive environment frame-by-frame based on your actions, rather than loading pre-built assets. Unlike a video game, Oasis 3 has no fixed map — it generates driving scenes in real time from text prompts, meaning any combination of weather, road, and conditions is possible without manual authoring.

### Who is Oasis 3 for right now?
Primarily autonomous-vehicle companies, which use it to simulate rare or dangerous driving scenarios at scale via API, per TechCrunch (10 June 2026). Decart plans to expand into robotics and other physical-AI applications once the driving domain is established.

### How much does Oasis 3 cost?
API access is priced at roughly $0.02 per second of simulated time, per TechCrunch (10 June 2026). That is cheap enough to generate thousands of scenario variants in a single overnight run.

### What are the real-world limitations testers found?
Multiple outlets flagged that the simulations are not perfectly accurate — 'photorealistic' does not mean 'physically correct'. The details varied by tester; the core concern is that sim-to-reality gaps in an AV training tool can teach the wrong behaviours, which surface only on real roads.

### Who are Decart's backers, and why does it matter?
Decart raised ~$300m at ~$4bn with Toyota, Adobe, eBay and Nvidia among backers, per TechCrunch (10 June 2026). Toyota's presence is strategically notable — an AV-adjacent manufacturer backing a world-model startup signals supply-chain intent, not just financial return.

## Sources

- [Decart's new world model can simulate hours of photorealistic driving — with some caveats](https://techcrunch.com/2026/06/10/decarts-new-world-model-can-simulate-hours-of-photorealistic-driving-with-some-caveats/) — TechCrunch, 2026-06-10
- [Decart Lays The Foundation For Physical AI Systems With Oasis 3](https://dataconomy.com/2026/06/10/decart-lays-the-foundation-for-physical-ai-systems-with-oasis-3/) — Dataconomy, 2026-06-10
- [Decart launches Oasis 3 world model for robotics and autonomous vehicle training](https://roboticsandautomationnews.com/2026/06/11/decarts-oasis-3-world-model-streams-realism-into-robotic-training-environments/102483/) — Robotics & Automation News, 2026-06-11