Space Bunny (anonymous M3.1 Flash) rebuilt a landing page from a 40-second screen recording
The author had stealth Space Bunny rebuild a full landing page from a 40-second screen recording, now confirmed as early M3.1 Flash, which MiniMax says scores 74% on DeepSWE and 91% on BabyVision at 3x speed and a third of M3's cost.
wait. I already tested this model and didn't even know it 💀 Last month a stealth model called Space Bunny rebuilt a whole landing page for me just by watching a 40 second screen recording. MiniMax just confirmed that was an early M3.1 Flash. Now it scores 74% on DeepSWE for coding and 91% on BabyVision for visual reasoning, which honestly makes sense after watching it rebuild a site from a video. And it's 3x faster and a third of the cost of M3. This matters because agents run hundreds of steps and every single step costs money. Strong AND cheap is what finally lets you leave one running all day.
This post quotes another post: x.com/1875078099538423808/status/211215086786854…