MiniMax M2 is an open-source, agent-native large language model (LLM) released by MiniMax on October 27, 2025. It is explicitly engineered for coding and agentic workflows (tool calling, multi-step automation), prioritizing low latency and cost-effective serving while delivering strong reasoning and tool-use capabilities.
Model Type: Chat
MiniMax today unveiled Music 1.5 (branded in some company channels as the Conch music model), a major upgrade to its generative-audio suite that the company says extends generation length and improves vocal realism while adding fine-grained, language-style control for creators. The release positions MiniMax to push AI music beyond short clips toward complete song production […]
On June 17, 2025, Shanghai-based AI leader MiniMax (also known as Xiyu Technology) officially released MiniMax-M1 (hereafter “M1”)—the world’s first open-weight, large-scale, hybrid-attention reasoning model. Combining a Mixture-of-Experts (MoE) architecture with an innovative Lightning Attention mechanism, M1 achieves industry-leading performance in productivity-oriented tasks, rivaling top closed-source systems while maintaining unparalleled cost-effectiveness. In this in-depth article, […]
MiniMax, a leading Chinese AI startup, has officially launched Hailuo 02, the latest iteration of its cutting-edge text-to-video engine This model sets a new benchmark in visual realism, physics simulation, and prompt precision. What Is Hailuo 02? Codenamed “Kangaroo”, Hailuo 02 succeeds Hailuo 01 and reportedly ranks #2 globally—behind ByteDance’s Seedance 1.0 but outperforming Google’s Veo 3. Utilizes a novel Noise‑aware […]
On June 17, Shanghai AI unicorn MiniMax officially open‑sourced MiniMax‑M1, the world’s first open‑weight large‑scale hybrid‑attention inference model. By combining a Mixture‑of‑Experts (MoE) architecture with the new Lightning Attention mechanism, MiniMax‑M1 delivers major gains in inference speed, ultra‑long context handling, and complex task performance . Background and Evolution Building upon the foundation of MiniMax-Text-01, which introduced […]
MiniMax Video-01 API provides developers with a comprehensive interface for harnessing sophisticated video analysis, transformation, and creation capabilities through a streamlined REST architecture, enabling seamless integration of advanced video intelligence into applications with minimal implementation complexity.
Model Type: Video
The Minimax ABAB7-Preview API is an advanced artificial intelligence framework that leverages deep learning, reinforcement learning, and multi-modal data processing to deliver scalable, real-time decision-making, natural language processing, and predictive analytics for complex applications.
Model Type: Video
