MiniMax M3.1 Flash Studio
A single flat-rate endpoint that handles agents, code, computer tasks and security reviews
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

Put MiniMax M3.1 Flash on long-horizon agent, coding and computer-use work: 200+ TPS output at roughly one-third of last-gen token spend.

All Tools

Discover our comprehensive AI-powered animation toolkit

Reasons Teams Standardize on MiniMax M3.1 Flash

Trained as the follow-up to the anonymous Space Bunny deployment, this Flash release has been measured on enterprise automation, software engineering, desktop operation and security testing.

  • Multi-Step Agent Automation
    Sits near the top of AutomationBench-AA and posts 76.2 on BankerToolBench, closing end-to-end SaaS runs in sales, marketing, operations, support, finance and HR.
  • Software Engineering Depth
    Hits 73.7% on DeepSWE v1.1 — reading repositories, adding features, patching defects and checking its own output with tests over long action chains.
  • Desktop Control and Security Audits
    Records 61.7 on OSWorld V2 and 29.6 on Agents' Last Exam for live-environment control, alongside 82% on CyberGym for finding and confirming vulnerabilities.

Running MiniMax M3.1 Flash in Three Moves

Go from first login to production-ready agent workloads with MiniMax M3.1 Flash in three short steps.

What MiniMax M3.1 Flash Delivers

Speed gains from deeper sparsity, predictable flat token rates and measured strength in automation, coding, desktop control and security work.

Deeper Sparsity, Quicker Replies

Extra sparsity across the serving stack pushes routine throughput past 200 TPS while the P50 stays above 170 TPS — close to three times the online inference speed of the prior release.

Predictable Per-Token Rates

At $0.15 per million input tokens, $0.60 per million output and $0.015 per million cached input, total agent budgets land about 70% under the earlier model.

SaaS Workflow Execution

A top AutomationBench-AA position and 76.2 on BankerToolBench come from multi-step planning, compliance handling and linked tool calls inside live SaaS environments.

Software Development on DeepSWE

73.7% on DeepSWE v1.1 places it among leading models for complex engineering that needs independent code reading, feature work, defect repair and test checks.

Computer Operation in Live Environments

Scores of 61.7 on OSWorld V2 and 29.6 on Agents' Last Exam, both leading among Chinese models, show it driving real environments, calling tools and finishing tasks.

Security Engineering Backup

An 82% CyberGym result, reinforced by in-house expert review, spans code auditing, patch work, pen testing, malware triage and reverse engineering without heavy spend.

FAQ

MiniMax M3.1 Flash: Common Questions

Straight answers on cost, speed, benchmark scores and how to get access to this Flash model.

1

So what is MiniMax M3.1 Flash, exactly?

It is MiniMax's latest Flash release, grown out of the anonymous Space Bunny deployment that handled more than 200 trillion tokens, with clear gains in agent work, coding, desktop control and security testing.

2

What does a million tokens cost?

Rates are flat: $0.15 per million input tokens, $0.60 per million output and $0.015 per million cached input, which brings typical agent runs roughly 70% below the previous model.

3

Where does the speed jump come from?

Added sparsity throughout the serving stack lifts routine throughput beyond 200 TPS, while the median, or P50, figure holds above 170 TPS.

4

Which benchmarks support these numbers?

A top AutomationBench-AA rank, 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?

Use the MiniMax API at platform.minimax.io, an M Plan subscription, or MiniMax Code, with the desktop build available from agent.minimax.io/download; open weights arrive later.

6

How does Flash compare with the Pro model?

Flash carries MiniMax's continuing work on balancing efficiency with capability, while Pro, focused on stretching raw intelligence further, lands at a later date.

Put MiniMax M3.1 Flash to Work

Sign in to the MiniMax API, choose an M Plan or install MiniMax Code — strong agent, coding, desktop and security results at about 70% less spend and close to three times the speed.