MiniMax

MiniMax is a Shanghai-based AI company building multimodal foundation models across text, image, video, audio, and music. Its MiniMax-M series LLMs target agentic and coding workflows, are open-weight, and are very cheaply priced, while the Hailuo AI product delivers top-tier AI video generation plus speech and music models. MiniMax listed on the Hong Kong Stock Exchange in January 2026 and appeals to cost-sensitive developers and creators, with data-residency caveats for Western enterprises.

MiniMax builds a broad multimodal stack spanning text, image, video, audio, and music. Its flagship LLM line evolved through the abab, M1, and M2 series, with recent models targeting tool-using agentic and coding workflows and offered as open weights. The consumer-facing Hailuo AI is a major AI video generation product, complemented by Speech-02 for voice and Music-01 for music. The M-series models are notably cheap and competitive on coding and agentic benchmarks, with the M2 line offering a large context window and the earlier M1 supporting ultra-long context up to a million tokens. The LLMs are published as open weights on Hugging Face and GitHub, while the video, speech, and music models are proprietary and hosted only. MiniMax listed on the Hong Kong Stock Exchange in January 2026, raising roughly US$619M with shares surging on debut. It had previously raised more than $850M from investors including Alibaba, Tencent, and miHoYo. Its low pricing and open-weight LLMs appeal to cost-sensitive global developers, while Hailuo has a large consumer and creator base, though data-residency and content-provenance concerns are worth weighing.