Space Bunny revealed as MiniMax M3.1 Flash: 62.2 agent score beats Opus 5
Space Bunny is confirmed as an early MiniMax M3.1 Flash build; the post cites a 62.2 agent benchmark versus Opus 5's 50.3 and GPT-5.6 Sol's 45.8, plus $0.15/$0.60 pricing and Oct 23 weights.
that “Space Bunny” model people ran 200 trillion tokens through? it was MiniMax M3.1 Flash all along. and for agent work, it’s 70% cheaper than the model before it while being nearly 3x faster. it’s built for actual end-to-end office work across business apps: sales, marketing, ops, support, finance, and HR. on the benchmark, it scored 62.2. Opus 5: 50.3 GPT-5.6 Sol: 45.8 pricing is just $0.15/M input tokens and $0.60/M output. so you’re getting top-tier agent performance at a fraction of the cost. it’s live today through the MiniMax API, MiniMax Code, and the M Plan with launch pricing. weights drop Oct. 23. i’ll be testing this one myself.
This post quotes another post: x.com/minimax_ai/status/2112150867868541221