Hboom

Hailuo 02 is MiniMax's video model — text-to-video and image-to-video, native 768p or 1080p at 6 or 10 seconds. It follows prompts closely and its motion physics hold together, which makes it handy for previz, concept work and short social clips.

An upgrade over Hailuo 02, still doing both text-to-video and image-to-video. Body movement, physics and facial micro-expressions are all better this round, and stylized looks — anime, ink-wash, game CG — hold up well enough for production video.

A sped-up cut of Hailuo 2.3, mostly for image-to-video — faster and cheaper per clip, aimed at short output. Good when you're batching a lot or working against the clock.

A lightweight MiniMax LLM — roughly 10B activated params — tuned for coding, agent workflows and modern app development. Cheap to run, and it still gets more done on real tasks.

A lightweight MiniMax LLM — roughly 10B activated params — tuned for coding, agent workflows and modern app development. Cheap to run, and it still gets more done on real tasks.

A MiniMax LLM pointed at real productivity work, trained across a range of messy digital work settings and carrying M2.1's coding ahead. It covers coding, agent tasks and everyday office workflows.

A MiniMax LLM pointed at real productivity work, trained across a range of messy digital work settings and carrying M2.1's coding ahead. It covers coding, agent tasks and everyday office workflows.

A MiniMax LLM pointed at real productivity work, trained across a range of messy digital work settings and carrying M2.1's coding ahead. It covers coding, agent tasks and everyday office workflows.

A MiniMax LLM pointed at real productivity work, trained across a range of messy digital work settings and carrying M2.1's coding ahead. It covers coding, agent tasks and everyday office workflows.

The next MiniMax LLM, built around getting real work done on its own. It handles multi-agent workflows and holds up on long, complex tasks that have to keep running.

The next MiniMax LLM, built around getting real work done on its own. It handles multi-agent workflows and holds up on long, complex tasks that have to keep running.

The next MiniMax LLM, built around getting real work done on its own. It handles multi-agent workflows and holds up on long, complex tasks that have to keep running.