Space Bunny（M3.1 Flash 匿名版）实测：两个硬 prompt 自测自修
作者在 OpenCode 用两个硬 prompt 测匿名 Space Bunny：写出带 alpha-beta 剪枝的四子棋引擎和监控面板，自己跑测试修 bug；官方版在 agent 类任务上更强。
last month i tested space bunny in opencode with two hard prompts it built a connect four engine with minimax, alpha-beta pruning and a live telemetry panel, then a full monitoring dashboard from an ascii wireframe. it wrote the files, ran its own tests and fixed its own bugs before handing them over that model is now revealed as MiniMax M3.1 Flash, and the official release is stronger on exactly this kind of agentic work - 62.8 on Automation-Bench, end-to-end workflows - 76.2 on BankerToolBench, multi-step planning and chained tool calls - 73.7% on DeepSWE v1.1, fixing and validating code through tests if space bunny impressed you, try the official version at http://agent.minimax.io
这条是引用着另一条写的：x.com/sakatayasha/status/2102969148934689207