minimax m3.1 flash Generator
A single flat-rate endpoint built for agent, coding, computer-use and security tasks
None

None

None

Long Story Video Skill

Turn your idea into 10s–10min video—AI writes scripts, prompts, and generates footage automatically.

Ads Video Skill

Ads Video Skill

Generate professional ads and sales videos—AI auto-generates scripts, prompts, and footage.

3D Science Explainer Video Skill

Convert scientific concepts into stunning 3D explain animations

AI Video Prompt Generator

Feedback

freeTrialImage.bannerPity

freeTrialImage.upgradeUnlock

  • ✓freeTrialImage.benefitHd
  • ✓freeTrialImage.benefitWatermark
  • ✓freeTrialImage.benefitUnlimited

minimax m3.1 flash

Explore minimax m3.1 flash for agent, coding and computer-use jobs — 200+ TPS output at a fraction of previous token spend.

All Tools

Discover our comprehensive AI-powered animation toolkit

Built for Real Work: minimax m3.1 flash

The successor to the anonymously run Space Bunny system, minimax m3.1 flash has been stress-tested across enterprise automation, software engineering, computer use and security operations.

  • End-to-End Agent Automation
    Places among the strongest entries on AutomationBench-AA and posts 76.2 on BankerToolBench, carrying SaaS workflows from sales and marketing through to finance, service, operations and HR until they are complete.
  • Software Engineering at Scale
    Scores 73.7% on DeepSWE v1.1 by reading existing code, shipping new features, squashing defects and running its own tests across extended step sequences.
  • Desktop Control and Security Audits
    Records 61.7 on OSWorld V2 and 29.6 on Agents' Last Exam for live-environment operation, plus 82% on CyberGym for finding and confirming vulnerabilities.

Getting Started with minimax m3.1 flash

Go from a first request to production-grade agent work in three short steps.

Core Strengths of minimax m3.1 flash

Quicker responses, predictable low token rates and measured strength across automation, software engineering, desktop control and security work.

Sparser Serving, Swifter Responses

Deeper sparsity in the serving stack pushes everyday output past 200 TPS while P50 holds above 170 TPS, making online inference close to three times quicker than the prior release.

Predictable Token Rates

$0.15 per 1M input tokens, $0.60 per 1M output tokens and $0.015 per 1M cached input bring total agent spend down roughly 70% versus the earlier model.

SaaS Workflow Completion

A leading AutomationBench-AA position and 76.2 on BankerToolBench reflect multi-step planning, financial compliance and chained tool calls running inside live SaaS systems.

DeepSWE Coding Performance

73.7% on DeepSWE v1.1 places it alongside top models on demanding software work that calls for independent code comprehension, feature builds, bug fixes and test validation.

Hands-On Computer Control

61.7 on OSWorld V2 and 29.6 on Agents' Last Exam — best among Chinese models on both — show it operating real environments, calling tools and finishing the job.

Security Engineering Backup

82% on CyberGym, confirmed by internal expert review, spans code auditing, vulnerability patching, penetration testing, malware analysis and reverse engineering at low cost.

FAQ

minimax m3.1 flash: Your Questions Answered

Quick answers on minimax m3.1 flash pricing, output speed, benchmark scores and how to get access.

1

So what is minimax m3.1 flash, exactly?

It is MiniMax's next Flash-generation model, grown out of the anonymously operated Space Bunny system that handled more than 200 trillion tokens; it moves ahead on agentic work, coding, computer use and cybersecurity.

2

What do a million tokens cost?

Rates stay flat: $0.15 per 1M input tokens, $0.60 per 1M output tokens and $0.015 per 1M cached input tokens, leaving overall agent usage about 70% cheaper than the model it replaces.

3

Why is inference nearly three times faster?

Greater sparsity throughout the serving stack: routine output reaches 200+ TPS while median (P50) speed stays above 170 TPS.

4

Which benchmark results support these numbers?

A top-tier AutomationBench-AA placement, 76.2 on BankerToolBench, 73.7% on DeepSWE v1.1, 61.7 on OSWorld V2, 29.6 on Agents' Last Exam and 82% on CyberGym.

5

How do I get access?

Via the MiniMax API at platform.minimax.io, an M Plan subscription or MiniMax Code, whose desktop build downloads from agent.minimax.io/download; weights arrive later.

6

How does Flash compare with the Pro model?

Flash continues MiniMax's exploration of the efficiency-capability balance, while Pro — aimed at pushing the boundaries of intelligence — lands later.

Put minimax m3.1 flash to Work Today

Call the MiniMax API, subscribe to an M Plan or install MiniMax Code to get agent, coding, computer-use and security results at about 70% less spend and close to three times the speed.