MiniMax M3.1 Flash
MiniMax's low-cost agentic and coding model, priced at $0.15/$0.60 per million tokens, officially rolled out Oct 19, 2026.
Overview
MiniMax officially rolled out M3.1 Flash on October 19, 2026, a text model aimed at agentic coding, computer use and security tasks; it had appeared earlier as M3.1-Flash-Preview inside MiniMax Code on September 27, 2026. It is available through the MiniMax API, the M Plan and MiniMax Code, priced at $0.15 per million input tokens, $0.60 per million output tokens and $0.015 per million cached input tokens, and MiniMax says everyday output reaches 200+ tokens per second (P50 170+). MiniMax reports 73.7% on DeepSWE v1.1, 61.7 on OSWorld V2, 29.6 on Agents' Last Exam, 76.2 on BankerToolBench and 82% on CyberGym; weights are not yet public ('will be released later'), and Vercel AI Gateway listed it as free to call on the same day.
This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.
Key features
- $0.15 input / $0.60 output / $0.015 cached input per million tokens
- 200+ tokens/second output in everyday use (P50 170+), per MiniMax
- 73.7% DeepSWE v1.1 and 61.7 OSWorld V2 (vendor-reported)
- Available via MiniMax API, M Plan, MiniMax Code and Vercel AI Gateway
- Open weights announced as coming later; not yet published
Use cases
- Low-cost agentic coding inside MiniMax Code or other agent harnesses
- Computer-use and browser automation agents
- Cybersecurity and financial-analysis agent tasks (CyberGym, BankerToolBench)
Pricing
$0.15 per 1M input tokens, $0.60 per 1M output tokens, $0.015 per 1M cached input tokens; also included in MiniMax M Plan subscriptions; free to call on Vercel AI Gateway at the time of listing.
Pricing reflects the source record and may change. Confirm current terms with the provider.
Dates and sources
- Released
- 19 Oct 2026
- Record updated
- 20 Oct 2026
- Record last checked
- 20 Oct 2026