Can Apertus 8B Instruct 2509 run on the Mac M3 Ultra 256GB?
Yes. At 8K context the best fit is F16, using about 16.7 GB of the 192 GB usable on the Mac M3 Ultra 256GB. Estimated speed ceiling: ~51 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|
| 2-bit (Q2_K) • | 4.2 GB Fits | 4.7 GB Fits | 7.7 GB Fits | 249 |
| 3-bit (Q3_K_M) • | 5.0 GB Fits | 5.5 GB Fits | 8.5 GB Fits | 197 |
| 4-bit (Q4_K_M) • | 5.9 GB Fits | 6.4 GB Fits | 9.4 GB Fits | 162 |
| 5-bit (Q5_K_M) • | 6.6 GB Fits | 7.1 GB Fits | 10.1 GB Fits | 141 |
| 6-bit (Q6_K) • | 7.3 GB Fits | 7.8 GB Fits | 10.8 GB Fits | 124 |
| 8-bit (Q8_0) • | 9.1 GB Fits | 9.6 GB Fits | 12.6 GB Fits | 96 |
| 16-bit (F16/BF16) • | 16.2 GB Fits | 16.7 GB Fits | 19.7 GB Fits | 51 |
• real GGUF file size from unsloth/Apertus-8B-Instruct-2509-GGUF. Speed is a bandwidth ceiling (819 GB/s), not a benchmark. Methodology.
More
Apertus 8B Instruct 2509 on every GPU and Mac · Every model that runs on the Mac M3 Ultra 256GB · Apple Silicon guide · Try other settings in the calculator
Data: Hugging Face Hub (swiss-ai/Apertus-8B-Instruct-2509), Apple specs. Data as of 2026-11-17.