MiniMax Launches Hailuo 2.3 Video Model with Improved Physics, Stylization, and Micro-Expressions
AI video model releases
MiniMax Launches Hailuo 2.3 Video Model with Improved Physics, Stylization, and Micro-Expressions
MiniMax released Hailuo 2.3, a video generation model built on the earlier Hailuo 02, focused on improving dynamic expression, character body movement, and camera motion realism.
The model shows better command-following and physics understanding, producing more fluid and precise character movements alongside near-photorealistic lighting, shadow, and color effects even with dynamic camera motion.
Hailuo 2.3 expands stylization support to include anime, illustration, ink wash painting, and game CG styles, building on the earlier "Live" model from Hailuo 01.
Live-action facial performance and micro-expression rendering is more natural in the new model, aimed at improving character acting quality.
During beta testing tied to the Double 11 shopping festival, creators using Hailuo 2.3 for e-commerce ads reported a higher success rate in generating usable high-quality content, reflecting improved object motion command response.