Can SmolLM3 3B Base run on the Mac M3 Ultra 256GB?
Yes. At 8K context the best fit is F16, using about 6.9 GB of the 192 GB usable on the Mac M3 Ultra 256GB. Estimated speed ceiling: ~133 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|
| 2-bit (Q2_K) • | 2.1 GB Fits | 2.4 GB Fits | 4.1 GB Fits | 653 |
| 3-bit (Q3_K_M) • | 2.4 GB Fits | 2.7 GB Fits | 4.4 GB Fits | 521 |
| 4-bit (Q4_K_M) • | 2.7 GB Fits | 3.0 GB Fits | 4.7 GB Fits | 428 |
| 5-bit (Q5_K_M) • | 3.0 GB Fits | 3.3 GB Fits | 5.0 GB Fits | 370 |
| 6-bit (Q6_K) • | 3.3 GB Fits | 3.6 GB Fits | 5.3 GB Fits | 324 |
| 8-bit (Q8_0) | 4.0 GB Fits | 4.3 GB Fits | 5.9 GB Fits | 251 |
| 16-bit (F16/BF16) | 6.7 GB Fits | 6.9 GB Fits | 8.6 GB Fits | 133 |
• real GGUF file size from mradermacher/SmolLM3-3B-Base-i1-GGUF. Speed is a bandwidth ceiling (819 GB/s), not a benchmark. Methodology.
More
SmolLM3 3B Base 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 (HuggingFaceTB/SmolLM3-3B-Base), Apple specs. Data as of 2026-11-17.