# MiniMax — MiniMax M3.1 Flash: Redefining Agent Performance and Cost

- Company: MiniMax (minimax.io)
- Announced: 2026-10-19
- Category: new-model
- Coverage: not counted
- Announcement: yes
- Group: models
- Source: https://www.minimax.io/blog/minimax-m31f
- Record: https://forck.live/items/24534-minimax-m3-1-flash-redefining-agent-performance-and-cost
- Subject: MiniMax models
- Models affected: MiniMax M3.1 Flash
- Pricing: Flat pricing per 1M tokens: $0.15 input · $0.60 output · $0.015 cached input

MiniMax has released MiniMax M3.1 Flash, a model it says makes online inference nearly three times faster and cuts usage costs by 70% compared with its predecessor. The company says the model improves on agentic tasks, coding, computer use and cybersecurity, and that an early version ran anonymously as Space Bunny, processing more than 200 trillion tokens. MiniMax reports 76.2 on BankerToolBench, 73.7% on DeepSWE v1.1, 61.7 points on OSWorld V2 and 29.6 on Agents' Last Exam, ranking first among Chinese models on the last two. It scored 82% on CyberGym, which MiniMax says shows leading capability in vulnerability discovery and validation. MiniMax attributes the speed to more aggressive sparsity, with output reaching 200+ tokens per second and median speeds consistently at 170+ TPS. Pricing is flat at $0.15 per million input tokens, $0.60 per million output tokens and $0.015 per million cached input tokens, and the model is available through the MiniMax API, M Plan and MiniMax Code, with weights to follow later.

## Evidence

Verbatim from https://www.minimax.io/blog/minimax-m31f:

> For agentic tasks, the model’s overall usage cost is 70% lower than its predecessor.
> 
> Flat pricing per 1M tokens: $0.15 input · $0.60 output · $0.015 cached input.

---

Record: https://forck.live/items/24534-minimax-m3-1-flash-redefining-agent-performance-and-cost
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md