All models
World modelsPreview
Decart
Oasis 3
Decart's action-conditioned world model for physical-AI training, generating synchronized, controllable simulation views in real time.
Plain-English overview
What Oasis 3 actually is
Oasis 3 is aimed at robotics and autonomous systems rather than entertainment-only worlds. A policy sends physical control signals, the generated environment responds, and the loop continues without rebuilding a hand-authored scene.
Decart reports under 200ms end-to-end and 22fps for three synchronized 512×768 views. Those are provider measurements for its infrastructure, not independent Cody benchmarks.
Good fit for
- Autonomous-vehicle edge-case simulation
- Closed-loop physical-AI training
- Teams exploring generated alternatives to fixed simulators
Category comparison
The facts that matter for world models
These are provider-published specifications, not Cody benchmark scores. Follow the linked sources for current limits and endpoint-specific exceptions.
- Price basisPublic API cost where available; research and private-beta models may not publish one.
- Not publicly listed
- Access stagePublic API, research experiment, partner onboarding, or private beta.
- API / partner access
- What it createsThe kind of interactive environment or embodied experience produced.
- Action-conditioned physical-AI simulator
- InteractionPublished frame rate, latency, or realtime behavior with its original basis.
- 22 fps; under 200ms end-to-end claim
- Visual outputProvider-published resolution or output format where available.
- 512×768×3 synchronized views
- API accessWhether developers can integrate the model today and under what access model.
- Yes, onboarding may be required
API and provider access
Where to get Oasis 3
Availability
Regions and access stage
API-accessible infrastructure focused first on autonomous-vehicle and physical-AI teams.
Public country and residency details are not listed; partnership access may be required for production use.
Check live availabilityData and training
The route matters.
The model page does not state a complete API training-use or retention policy. Physical-AI teams should obtain contractual data-handling terms before contributing proprietary scenarios or fleet data.
This is a concise reading of the cited provider material, not legal advice. A third-party gateway can have different storage, routing, training, and residency terms from the model maker's direct API.
Read the provider policyFrequently asked questions
Oasis 3 FAQ
What is Oasis 3?
Decart's action-conditioned world model for physical-AI training, generating synchronized, controllable simulation views in real time. Oasis 3 is aimed at robotics and autonomous systems rather than entertainment-only worlds. A policy sends physical control signals, the generated environment responds, and the loop continues without rebuilding a hand-authored scene.
When was Oasis 3 released?
The provider does not publish a clear release date for Oasis 3 in the source material reviewed by Cody.
Where can I access Oasis 3?
API-accessible infrastructure focused first on autonomous-vehicle and physical-AI teams. The access routes listed in this guide are Decart.
How much does Oasis 3 cost?
Not publicly listed. Decart offers API and partnership access but does not publish a self-serve price on the model page.
Where is Oasis 3 available?
API-accessible infrastructure focused first on autonomous-vehicle and physical-AI teams. Public country and residency details are not listed; partnership access may be required for production use.
Is my Oasis 3 API data used for training?
The model page does not state a complete API training-use or retention policy. Physical-AI teams should obtain contractual data-handling terms before contributing proprietary scenarios or fleet data. The policy belongs to the provider route and account terms, so verify it again before production use.
Research trail
Sources for this guide
Keep comparing
More world models models
Google DeepMind · Research access
Genie 3
Google DeepMind's research world model for generating photorealistic, controllable environments that can be explored in real time from a text description.
Read guideWorld Labs · Stable
Marble 1.1 Plus
World Labs' public API model for turning text, images, panoramas, multi-view sets, or video into large explorable 3D worlds and spatial assets.
Read guideOdyssey · Private beta
Odyssey-2 Max
Odyssey's largest general world model, focused on real-time next-state simulation, physical accuracy, open-ended futures, and longer interactive rollouts.
Read guide