Tiny 540K-Parameter Robot Policy Hits 95% Success

The MINERVA paper shows a vision-language-action policy with just 540K parameters achieving 95.1% success on LIBERO, using 7700x fewer parameters than π0.5. Meanwhile, a Meta AI researcher cautions that open-source VLA models still struggle with basic representation alignment and have not reached a true 'GPT-3 moment.'

2026-09-09 ~ 2026-09-09 · 2 related posts

• 0.54M-parameter robot policy hits 95% on LIBERO, 7700x smaller than π0.5 — ChongZzZhang · 2026-09-09
• Open-source VLAs still struggle with basic representation alignment, researcher argues — ChongZzZhang · 2026-09-09