KOSUNI CRYPTO
Korea's Crypto Pulse, in English
$BTC $ETH $XRP $SOL $DOGE
News
Neutral

MiniMax Launches M3.1 Flash, Claims 70% Lower Agent Work Costs

Published October 20, 2026 3:33 AM · 4 views
MiniMax Launches M3.1 Flash, Claims 70% Lower Agent Work Costs

Chinese AI company MiniMax has released its new model M3.1 Flash. On October 20, the company announced on its official blog that agent work usage costs have dropped by 70% compared to previous models, with online inference speeds approximately three times faster. MiniMax stated that M3.1 Flash's output speed exceeds 200 tokens per second. API pricing is set at $0.15 (about 200 KRW) for input and $0.6 (about 800 KRW) for output per 1 million tokens.

Regarding performance, MiniMax reported that M3.1 Flash achieved a score of 61.7 on the OSWorld V2 benchmark, ranking first among Chinese models. The company also explained that 'Space Bunny,' an anonymous model that processed over 20 trillion tokens, was the initial version of M3.1 Flash. M3.1 Flash is available through MiniMax API, M Plan, and MiniMax Code. MiniMax plans to open-source the model weights later.

Korean Source

This article is an English localization of a Korean-language crypto news report. Original headline: 미니맥스, M3.1 Flash 출시…"에이전트 작업 비용 전작 대비 70% 절감"