MiniMax: MiniMax M2.1

MiniMax-M2.1 is a cutting-edge large language model, specifically engineered for coding, agentic workflows, and modern application development. Despite having only 10 billion activated parameters, it delivers a significant leap in real-world capability, maintaining exceptional latency, scalability, and cost efficiency. This model is ideal for developers and businesses seeking high performance without compromising on speed or budget. Compared to its predecessor, M2.1 provides cleaner, more concise outputs and faster perceived response times. It excels in multilingual coding across major systems and application languages, achieving 49.4% on Multi-SWE-Bench and 72.5% on SWE-Bench Multilingual. It also serves as a versatile agent "brain" for IDEs, coding tools, and general-purpose assistance. With a Context Window of 204K tokens and Max Output of 131K tokens, and pricing at $0.27/$1.10 per 1M tokens (input/output), MiniMax M2.1 is a powerful and cost-effective solution. It supports streaming, functions, and long_context capabilities. To optimize performance, MiniMax recommends preserving reasoning between turns.