SignalTech/AI
MiniMax M3.1 Flash
High-priority alert based on source significance and keyword relevance.
Context
What happened: MiniMax rolled out MiniMax M3.1 Flash, a next-generation Flash model. An early version ran anonymously as Space Bunny, processing more than 200 trillion tokens.
Key numbers: Online inference nearly three times faster; usage costs down 70%. 200+ TPS output, 170+ TPS median.
Why it matters: MiniMax says the model reaches the level of leading models on agentic tasks, coding, computer use and cybersecurity at a very low task cost. It is available via the MiniMax API, M Plan and MiniMax Code, with weights to be released later.
Coverage: 6 outlets in 24h
What changed
MiniMax rolled out MiniMax M3.1 Flash, a next-generation Flash model. An early version ran anonymously as Space Bunny, processing more than 200 trillion tokens.
Why it matters
MiniMax says the model reaches the level of leading models on agentic tasks, coding, computer use and cybersecurity at a very low task cost. It is available via the MiniMax API, M Plan and MiniMax Code, with weights to be released later.