IT之家 · ITHome
MiniMax M3.1 Flash launches; developer reports faster inference and 70% lower Agentic costs
MiniMax M3.1 Flash has launched, with its developer reporting a nearly threefold increase in online inference speed and 70% lower overall costs for Agentic tasks versus its predecessor. The company also reports gains in coding, computer use and cybersecurity. The model is available through MiniMax API, MPlan and MiniMax Code; model weights are planned for release later.
What to know
- 01
Listed prices per million tokens are 1 yuan for input, 4 yuan for output and 0.1 yuan for cache reads.
- 02
The company reports scores of 73.7% on DeepSWE v1.1 and 82% on CyberGym.
Why it matters
The company-reported performance and pricing offer a reference for comparing model capabilities and usage costs in Agentic tasks and software development.
AI-assisted brief. Read the source for full details. Editorial & sourcing policy
