MiniMax M2.5 Uses 10B Active Parameters per Token, Aiming for Cheaper Always-On Agents

MiniMax M2.5 hits about 80% on Sweetbench and runs near 100 tokens per second, helping teams deploy faster models on tighter budgets.

China’s MiniMax releases cheap AI model 'designed for real-world productivity'

Company says it is 'first frontier model ... delivering on the promise of intelligence too cheap to meter' Chinese artificial ...
TechBooky

MiniMax Unveils M2.5 Models to Cut Frontier AI Costs

Chinese AI startup MiniMax, based in Shanghai, has introduced a new M2.5 language model in two variants, positioning it as ...
Hosted on MSN

China’s MiniMax debuts M1 AI model that it says costs 200x less to train than OpenAI’s GPT-4

It’s becoming a familiar pattern: Every few months, an AI lab in China that most people in the U.S. have never heard of releases an AI model that upends conventional wisdom about the cost of training ...
SiliconANGLE

MiniMax releases M2.1 AI model for multi-language programming versatility

Chinese artificial intelligence startup MiniMax today announced the release of M2.1, a significantly enhanced performance for real-world complex tasks and agentic capabilities across more programming ...
AOL

Disney, Warner Bros. Discovery, NBCU Sue Chinese AI Company MiniMax, Alleging It ‘Pirates and Plunders’ Studios’ Copyrighted Works on ‘Massive Scale’

Disney, Warner Bros. Discovery and NBCUniversal are taking legal aim at another AI company that they allege is massively ripping off their intellectual property. On Tuesday, the three media companies ...
TechNode

MiniMax CEO predicts another 10x drop in AI inference costs within two years

MiniMax CEO Yan Junjie said inference costs for leading AI models could fall by another order of magnitude within one to two years, following a similar drop over the past year. Speaking at the 2025 ...