MiniMax Rolls Out M3.1 Flash With Faster Inference and 70% Lower Cost
MiniMax has officially launched M3.1 Flash, a next-generation Flash model that targets agentic tasks, coding, computer use and cybersecurity. The company says inference is nearly three times faster and usage costs are 70% lower.
MiniMax launches M3.1 Flash
MiniMax has officially rolled out MiniMax M3.1 Flash, describing it as a next-generation Flash model with a generational leap in agentic tasks, coding, Computer Use and cybersecurity. The company says online inference is nearly three times faster and usage costs are reduced by 70%.
Space Bunny, which processed more than 200 trillion tokens while operating anonymously, was an early version of M3.1 Flash. The model is now available through the MiniMax API, M Plan and MiniMax Code. Model weights will be released later.
Agentic, coding and security benchmarks
MiniMax says M3.1 Flash reaches the level of leading models in complex agentic tasks, computer use, software development and cybersecurity. It ranks among the top performers on AutomationBench-AA and can complete real enterprise workflows across SaaS systems, covering sales, marketing, operations, customer service, finance and HR end to end.
The model scores 76.2 on BankerToolBench, with strengths in multi-step planning, consistency across documents and files, financial compliance and chained tool calls. On coding, it achieves 73.7% on DeepSWE v1.1, which tests complex software work such as understanding code, developing features, fixing issues and validating results through testing.
For computer use, M3.1 Flash scores 61.7 points on OSWorld V2 and 29.6 points on Agents' Last Exam. MiniMax says it ranks first among Chinese models on both benchmarks. In cybersecurity, it achieves 82% on CyberGym for vulnerability discovery and validation, and internal security experts evaluated it in code auditing, vulnerability patching, penetration testing, malware analysis and reverse engineering.
Faster inference and pricing
MiniMax optimized M3.1 Flash inference efficiency through more aggressive sparsity. In everyday use, output speeds can reach 200+ tokens per second, with median P50 speeds consistently reaching 170+ tokens per second.
For agentic tasks, the overall usage cost is 70% lower than its predecessor. Flat pricing per 1M tokens is $0.15 for input, $0.60 for output and $0.015 for cached input.
MiniMax positions M3.1 Flash as a balance between efficiency and capability. The company says Pro, which pushes the boundaries of intelligence, will also arrive later.
SiTech — AI-powered web development
We build fast, modern websites and bring AI into real business workflows. Have a project or a question? We'd love to help.