MiniMax M2.7 on Vercel AI Gateway

MiniMax M2.7 is now live on Vercel AI Gateway, available in two deployment options: standard and high-speed variants. This release represents a significant advancement in the M2-series lineup, with notable improvements across software engineering, agentic workflows, and professional office tasks.

Key Capabilities

The M2.7 model introduces native multi-agent collaboration and complex skill orchestration, enabling developers to build sophisticated agentic systems. The model features:

• Native support for multi-agent collaboration and dynamic tool search
• Enhanced production debugging and end-to-end project delivery capabilities
• Improved performance on software engineering and workflow automation tasks

High-Speed Variant

The high-speed variant (minimax/minimax-m2.7-highspeed) delivers the same quality output at approximately 100 tokens per second, optimized for latency-sensitive applications. This variant costs 2x the standard rate but provides significant speed improvements for real-time use cases.

Getting Started

Developers can access M2.7 through the Vercel AI SDK by setting the model parameter to minimax/minimax-m2.7 (standard) or minimax/minimax-m2.7-highspeed. AI Gateway provides a unified API surface with built-in observability, usage tracking, cost management, and intelligent provider routing.

Integration Benefits

Vercel AI Gateway offers additional features including Bring Your Own Key (BYOK) support, automatic retries, failover strategies, and performance optimizations designed to achieve higher-than-provider uptime. Developers can monitor performance, optimize costs, and configure routing policies through a single interface.