Mumbai, May 28 -- Artificial intelligence software developer Helm.ai has launched two foundation models, GenSim-3 and VidGen-3, establishing a native Full HD (1920x1080) resolution standard for generative simulation across a 6-camera, 360-degree surround-view suite.

The architecture delivers 5x the pixel density of industry benchmarks to assist automotive developers facing the limitation where physical collection of edge cases becomes logistically restrictive.

Traditional generative world models typically cap resolution at roughly 0.4 megapixels per camera. Helm.ai's platform outputs a native 2 megapixels per camera, yielding a synchronised 12-megapixel synthetic canvas per timestep. This specification matches the hardware parameters of...