MiniMax M3

MiniMax M3 is Shanghai-based MiniMax’s frontier open-weight release, combining a 1M-token context window with native multimodal input — text, image, and video — aimed at full-repository code understanding rather than file-by-file work.

Why it stands out

• 1,048,576-token context window, with a guaranteed minimum of 512K even under high load.
• Native multimodal input across text, image, and video in the same coding session.
• Open-weight, giving teams the option to self-host instead of depending on a hosted API.
• Tiered pricing by input length, keeping typical coding and conversation calls cheap while reserving a higher rate for ultra-long, full-repo parsing.

Good for

Teams working across large, multimodal codebases who want frontier-tier long-context performance with open weights.