How is it that these models boast these amazing benchmark results, but using it ... | Hacker News

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

tjwebbnorfolk on April 29, 2025 | parent | context | favorite | on: Qwen3: Think deeper, act faster

How is it that these models boast these amazing benchmark results, but using it for 30 seconds it feels way worse than Gemma3?

manmal on April 29, 2025 [–]

Are you running the full versions, or quantized? Some models just don't quantize well.

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