minimax_api_key enables all capabilities.
Text Chat
MiniMax-M3.1-Flash always thinks (thinking cannot be disabled); tune its depth with reasoning_effort: low / medium / high / xhigh / max (default max). MiniMax-M3 still supports disabling thinking. See the MiniMax API reference.
Image Understanding
MiniMax-M3.1-Flash and MiniMax-M3 natively accept text, image, and video input through the OpenAI-compatible chat endpoint. Onceminimax_api_key is configured, the Agent’s Vision tool automatically uses a MiniMax multimodal model for image requests, with no need to specify a separate vision model.
Image Generation
image-01.
Text-to-Speech (TTS)
Common voice examples:
For the full voice list (70+ voices across Chinese / Cantonese / English / Japanese / Korean), see the system voice list, or select visually in the Web Console under “Model Management → Text-to-Speech”.
