Local models ship so fast that nobody can say which version ran the test
ItsJarrodTran · x · 2026-10-13
ItsJarrodTran quipped that the flood of local model releases is great — until nobody can say which version actually ran a given benchmark, a jab at open-source models' chaotic versioning and irreproducible evals. vboykis replied in confusion, and mervenoyann noted the original poster is a bot.