Hailuo 2.3

High-fidelity video generation for realistic human motion and cinematic VFX with strong prompt and style adherence in text-to-video and image-to-video workflows.

Overview

MiniMax-Hailuo-2.3 is a video generation model designed to produce high-quality videos featuring realistic human motion, cinematic visual effects, and expressive characters. It supports both text-to-video and image-to-video workflows, accommodating detailed prompts and first-frame images to guide generation. The model offers resolutions of 768p and 1080p, with 1080p limited to 6-second videos and 768p supporting up to 10 seconds.

Strengths / What it does well

• Generates visually consistent, high-fidelity cinematic animations
• Excels at animating realistic human movements and complex visual effects
• Strong adherence to textual prompts and style directives across workflows
• Supports flexible aspect ratios by default for text-to-video (16:9) and based on source image for image-to-video

Limitations

• Does not support last frame preservation in output videos
• 1080p resolution is limited to short video duration (6 seconds)

Best use cases

Ideal for cinematic workflows requiring realistic and expressive character animation, stylistic visual effects, and high-detail scene generation from both text prompts and images. Suitable for content creators who need high visual quality and versatile input modes in video generation.