Not quite true. Depends on number of KV heads. GLM4 32b at IQ4 quant and Q8 cont... | Hacker News

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

iamnotagenius on Aug 7, 2025 | parent | context | favorite | on: Qwen3-4B-Thinking-2507

Not quite true. Depends on number of KV heads. GLM4 32b at IQ4 quant and Q8 context can run full context with only 20GiB VRAM.

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