AI Model

MiniMax M2 M2.7 Highspeed

Released on March 18, 2026

Changes from previous version

Same intelligence target as M2.7 with a high-throughput serving profile for latency-sensitive deployments.

Release Summary

High-TPS variant of MiniMax-M2.7 with identical results to M2.7 at faster speed for agent and coding APIs.

Timeline

March 18, 2026

MiniMax M2.7 Highspeed released

MiniMax adds M2.7-highspeed next to the standard M2.7 API endpoints.