Odyssey-3 is a new generative world model that you can try for free
Odyssey opened a free preview of Odyssey-3, a 14-billion-parameter world model for real-time interactive environments.
California-based Odyssey opened a public research preview of Odyssey-3, a generative world model that turns text prompts into interactive environments users can explore in real time. The base model has 14 billion parameters and generates 832×480 video with an autoregressive diffusion transformer; Odyssey-3 Pro targets 1280×720. The company reports a 66.1 Physics-IQ score, but that figure is from a single selected-video run and does not meet the benchmark’s four-run protocol; the four-run average was 63.37. Odyssey also describes pairing the model with controllers for robotic arms, drones, and video-game agents, and said it raised $310 million in June 2026.
- Odyssey-3 public research preview generates interactive worlds from text in real time.
- Base model has 14 billion parameters at 832×480; Odyssey-3 Pro targets 720p.
- Claimed Physics-IQ score of 66.1 used selected videos and missed the four-run protocol.
- Controllers on Odyssey-3 steered robot arms, drones, and transferred skills across games.
- Odyssey raised $310 million in June 2026, including Amazon and AMD Ventures.
- Vendors
- Odyssey
- Organizations
- OdysseyFlexionGoogle DeepMindWorld LabsAmazonAMD
- Countries
- United StatesSwitzerland
From the original article
4 min read at the-decoder.com
Odyssey is making its world model Odyssey-3 available as a public research preview. The 14-billion-parameter model generates interactive environments from text prompts in real time and can reportedly control robots, drones, and even GTA V. Odyssey claims top scores on physics benchmarks, though the methodology behind those results has limits.
By Tomislav Bezmalinović · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.