MiniMax Debuts M3.1 Flash, a High-Speed Model for AI Agents at 70% Lower Cost

MiniMax launched M3.1 Flash on October 19, 2026, a streamlined version of its flagship model that triples inference speed while cutting operational costs by 70% compared to its predecessor. The release emerged from Space Bunny, an anonymous stealth model that processed over 200 trillion tokens on OpenRouter before its public reveal, ranking first among comparable models for 10 consecutive days during its incognito run.