22 signals · 0 breaking · last activity 3MO AGO
A new-generation open general-purpose multimodal video model. It understands creative intent across multimodal context — text, image, video, and audio — and delivers more natural, coherent generation and expression.
Experience the New Music Generation Capabilities
Released MiniMax-M3, the latest M-series language model for agentic reasoning, tool use, coding, multimodal chat input, and long-context tasks.
Released MiniMax-M2.7 series (MiniMax-M2.7 / M2.7-highspeed). Beginning the journey of recursive self-improvement.
Released Music2.5+, Instrumental Unlocked, break through style boundaries
Released MiniMax-M2.5 series models (MiniMax-M2.5 / M2.5-highspeed). Achieving or setting new SOTA benchmarks in programming, tool calling and search, office productivity and other scenarios.
Natural Sound Tags, lifelike voice, pristine audio quality
Released Music 2.5, Direct the Detail. Define the Real.
Released MiniMax-M2.1. Polyglot Programming Mastery, Precision Code Refactoring.
Released Music 2.0, enhanced musical expression that makes every melody deeply emotional.
Released Speech 2.6 — More human and instantly responsive voice for your agent.
Released MiniMax-Hailuo-2.3 & 2.3-Fast, featuring breakthroughs in body movement, facial expressions, physical realism, and prompt adherence.
MiniMax-M2 Released. An Efficient Model for the Agentic Era. Get free MiniMax M2 model API calls now. Offer ends November 7, 2025, at 00:00 UTC.
New music model released. Now supports 4-minute songs, getting back to the core of great sound.
Released a next-generation speech synthesis model featuring broader language coverage and exceptionally high voice similarity.
Released Music-1.5 (Beta), a next-generation music generation model that supports creating music from musical prompts and lyrics.
Released MiniMax Hailuo 02, a next-generation video generation model with support for 1080p resolution and video generation of up to 10 seconds.
Released the Speech-02 model series, delivering hyper-realistic vocal performance. The models are distinguished by exceptional prosody and stability, along with outstanding voice similarity and audio quality.
Released the Image-01 model, which supports text-to-image generation in multiple sizes.
Released a video generation model with enhanced camera control, offering superior adherence to camera movement directives and enabling cinematic storytelling through its shot composition.
Released the next-generation LLM, MiniMax-Text-01, and the vision-language model, MiniMax-VL-01.