🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the Apple M3 Max 36GB run DeepSeek-R1-Distill-Qwen-1.5B?
Yes — here's how fast and up to what context, with real measured numbers.
DeepSeek-R1-Distill-Qwen-1.5B on the Apple M3 Max 36GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 417k | 57.3-69.1 t/s | ~1.9-2.3 (slow) |
| Q8 | yes | 444k | 114.6-138.2 t/s | ~3.8-4.6 (slow) |
| Q4 | yes | 456k | 204.7-246.8 t/s | ~6.8-8.2 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
FAQ
Can the Apple M3 Max 36GB run DeepSeek-R1-Distill-Qwen-1.5B?▶
Yes. At Q4 it generates ~204.7-246.8 tok/s and fits up to 456k context; beyond that it spills to SSD swap and slows to ~6.8-8.2 tok/s.
How much context fits?▶
At Q4, up to about 456k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.