Not sure "decimated" is a fitting word for "slightly higher performance on some ... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

mupuff1234 on April 17, 2025 | parent | context | favorite | on: Gemini 2.5 Flash

Not sure "decimated" is a fitting word for "slightly higher performance on some benchmarks".

fwip on April 17, 2025 | [–]

Perhaps they were using the original meaning of "one-tenth destroyed." :P

ein0p on April 17, 2025 | [–]

66.8% error rate reduction for o4-mini on AIME2025, and 21% error rate reduction on MMMU isn't "slightly higher". It'll be quite noticeable in practice.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact