I am kind of becoming a Meta fanboy. No marketing fluff No unnecessary drama Only shipping! Muse Spark 1.3 is beating GPT 5.6 and Opus 5 at their own game. → DeepSWE v1.1: 75.4, beats GPT 5.6 Sol (73.0) and Opus 5 (74.0) → Codebase understanding (SWEAtlas): 59.4, clears both by 6+ points → Terminal coding (Terminal-Bench 2.1): 88.8, ties GPT, beats Opus's 86 I didn’t see that coming.