Benchmarking Fable, Sol, and Kimi K3 on SlopCodeBench

1. If we're using native harnesses, I'd have preferred you use kimi code, not opencode 2. The variation in the two kimi providers just shows how you can't trust n = 1 trials