Decision Index 0.4 switches to private benchmarks to fight benchmark gaming

Hugging Face's multimodalart updated Decision Index to version 0.4, scoring 224 models entirely on private and internal benchmarks to counter benchmark gaming.

2026-10-22 ~ 2026-10-22 · 2 related posts

• Decision Index 0.4: 224 models scored on private benchmarks only to curb benchmaxxing — multimodalart · 2026-10-22

1 near-duplicate retellings: Thom_Wolf