Genie Models by Google DeepMind

Google DeepMindProprietary
Google DeepMind releases · 53 in the last 12 months · this family litChangelog →
2 models2025–2026

Last refreshed 2026-11-18. Next refresh: weekly.

Details

ResearcherGoogle DeepMind
LicenseProprietary
Commercial useRestricted
Models2
Released2025–2026

Capabilities

VisionAll models
MultimodalAll models

Links

Website

About

Genie is Google DeepMind's line of interactive world models. It includes Genie 3, announced August 5, 2025 and later offered in the Project Genie Labs prototype, and Genie 4, announced October 27, 2026 and made available in Project Genie on November 18, 2026.

Decision facts

Best fit
video generationvision and multimodal work
Capability starting point
Genie 4 with multimodal inputs
Lowest tracked input
Not tracked
Closest related family
DiffusionGemma

Current Variants

Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.

2 in view
Genie 4Current

Use when the workload needs video generation, multimodal inputs, and audio.

2026-10video generationmultimodal inputsaudio
Genie 3Current

Use when the workload needs video generation and multimodal inputs.

2025-08video generationmultimodal inputs

Release Timeline

2 release groups
2026-10
1 current
Genie 4
video generationmultimodal inputsaudio
Current
2025-08
1 current
Genie 3
video generationmultimodal inputs
Current

Specifications(2 models)

Genie model specifications comparison
ModelReleasedVisionMultimodal
Genie 42026-10YesYes
Genie 32025-08YesYes