EXPLORINEER
vi

Lưu trữ — CN — technode

MiniMax releases M2 open-source model , offering double speed at 8% of ClaudeSonnet’s price

MiniMax, a China-based AI startup, has launched its open-source model, MiniMax M2, designed for agent workflows and coding tasks. The M2 offers double the inference speed of Anthropic's Claude Sonnet 4.5 at only 8% of its cost. It's free for a limited time and supports various platforms. The model is available on Hugging Face along with a dedicated agent platform.

Điểm chính

  • MiniMax has released M2, an open-source large language model.
  • M2 is specifically focused on agent workflows and coding tasks.
  • It provides twice the inference speed of Claude Sonnet 4.5.
  • M2 costs only 8% of Claude Sonnet's API price.
  • The model is currently free to use for a limited time.
  • MiniMax M2 shows strong performance in code generation and tool use.
  • The weights for M2 are available on Hugging Face.
  • MiniMax supports deployment with vLLM and SGLang.
  • MiniMax offers its M2-powered agent platform with features for automation.

Vì sao quan trọng

  • The release of MiniMax M2 aligns with the trend towards open-source AI models gaining popularity.
  • In 2025, AI models aim for cost-effectiveness and high performance to compete with established players.
  • Open-source models like MiniMax M2 could disrupt current market leaders by providing accessible alternatives.

MiniMax M2's launch signifies a significant advancement in open-source AI, providing high performance at a low cost, which could reshape the competitive landscape in AI technologies.