Crypto news from Coin World: According to Beating's monitoring, the global model company Odyssey has released Odyssey-2 Max, featuring three times the parameters of its predecessor Odyssey-2 Pro and ten times the training compute power. This is the company's largest general-purpose world model to date. Unlike bidirectional video models like Sora and Veo, Odyssey-2 Max adopts a causal autoregressive architecture, predicting the next state frame by frame and accepting real-time interactive inputs, capable of generating over 120 seconds of simulation visuals without crashing. In terms of physical simulation accuracy, Odyssey-2 Max scored 58.52 in the VBench 2 physics subcategory, outperforming Odyssey-2 Pro's score of 49.67 and NVIDIA's Cosmos-Predict2.5-14B at 44.92; it scored 93.02 in the PAI-Bench physics subcategory. The model is based on an autoregressive diffusion Transformer (AR DiT) architecture, utilizing continuous flow matching for generation, and achieving real-time inference through stepwise denoising distillation. Training was completed on hundreds of NVIDIA B200 GPUs, conducted in three phases: large-scale video pre-training, interaction and task conditioning, and long-sequence stability training. Odyssey was founded by Oliver Cameron (CEO) and Jeff Hawke (CTO), both from the autonomous driving industry, and has raised a total of $27 million to date. Odyssey-2 Max is now available for Private Beta to partners in the fields of robotics, gaming, simulation, defense, and interactive systems.