3 ms·
>Opus 4.6 Max self-hosted at 30 tok/s on a 5k Macbook in Aug 2026. The LLM timelines are crazy. How much memory does this translate to and what quantization (i
by gruez 1mo ago
>Opus 4.6 Max self-hosted at 30 tok/s on a 5k Macbook in Aug 2026. The LLM timelines are crazy.
How much memory does this translate to and what quantization (if any) were applied?
- rohansood15 1mo ago128GB, 4-bit quantized.