MiniMax launches the M3.1 Flash multimodal MoE agentic model via its API
Model release Provisional 78% confidence first seen
MiniMax has launched its M3.1 Flash multimodal Mixture-of-Experts (MoE) model for agent-style tasks through its API and subscriptions. The company announced pricing and performance targets, including lower per-token costs, and said model weights are scheduled to arrive later while reasoning is enabled with integration updates.
M3.1 Flash
MiniMax
MiniMax Group Inc.
MiniMax M3.1 Flash
M3
M3 large language model
SiliconANGLE
Space Bunny