MiniMax just halved its token prices. Your AI agent bill may not follow.
Summary
MiniMax launched M3.1 Flash and cut its token prices by half. The company claims the model lowers agent costs by 70 percent, but the article notes that real savings depend on workload, token usage, and how many loops an agent needs. Independent benchmarks show mixed results: the model looks competitive on coding tasks, weaker on computer use, and strong on some factual benchmarks but with a high hallucination rate. The article also highlights its one-million-token context window, upcoming weight release, and the fact that the license has not yet been published.
AskAI Classifications
Sectors
No sectors detected
Functions
Developer and IT Infrastructure
Development Platforms
Linked Companies
MiniMax
$25M to $50M