MiniMax Ships Hailuo 2.3 Video Generator with Speed and Quality Upgrades

Chinese AI company MiniMax has released Hailuo 2.3 and Hailuo 2.3 Fast, the latest versions of its text-to-video generation platform. The update brings improvements to video quality and generation speed, continuing the company’s push into the competitive AI video space.

MiniMax announced the release as part of its ongoing development of the Hailuo platform, which competes with tools like Runway and Pika Labs in the text-to-video generation market. The dual-model approach offers users a choice between quality and speed depending on their needs.

Two Models for Different Use Cases

The release includes two distinct versions: Hailuo 2.3 focuses on video quality improvements, while Hailuo 2.3 Fast prioritizes generation speed. This mirrors the approach taken by other AI companies that offer multiple model variants to serve different user priorities.

The standard 2.3 model targets users who need higher-quality output and can wait longer for results, while the Fast variant serves those who need quicker turnaround times for rapid prototyping or high-volume content creation.

Quality and Performance Improvements

MiniMax claims the 2.3 models deliver better video coherence and visual fidelity compared to previous versions. The improvements reportedly include more stable motion tracking, better object consistency across frames, and enhanced detail preservation in generated videos.

The Fast model maintains competitive quality while reducing generation times, though specific performance metrics weren’t provided in the announcement. This balance between speed and quality has become a key differentiator in the text-to-video space as the technology matures.

Platform Access and Availability

Both models are available through MiniMax’s web platform, continuing the company’s direct-to-consumer approach rather than API-first distribution. Users can access the updated models through their existing Hailuo accounts without additional setup requirements.

The release maintains MiniMax’s position in the Chinese AI market while competing internationally with established players. The company has been building its video generation capabilities as part of a broader AI platform strategy.

Bottom Line

Hailuo 2.3 represents incremental progress in text-to-video generation rather than a breakthrough moment. The dual-model approach is practical for users with different speed versus quality priorities, but without detailed benchmarks, it’s hard to assess how these improvements stack up against competitors like Runway Gen-3 or OpenAI’s Sora. The release keeps MiniMax competitive in a rapidly evolving space where consistent improvements matter more than occasional leaps.