GPUs & Macs › Apple M1 Ultra
Apple M1 Ultra: local LLM speed
1 measured result for the Apple M1 Ultra.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Gemma 4 E4B | 69.7 | 1,173 | Q4_K_M | Metal | tg128 | majid.info log details |
What fits on the Apple M1 Ultra
Estimates for this chip are scaled from a measured run: Gemma 4 E4B Q4_K_M at 69.7 tok/s (Metal, source). Fit assumes the whole model in GPU memory with ~8K context.
With 16 GB unified memory
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| gemma-4-12B-it | Q6_K | 9.8 GB | 35.5 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 36.4 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 77.4 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 36.4 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 81.1 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 51.7 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 289 estimate |
| gemma-4-E2B-it | Q8_0 | 5.0 GB | 69.9 estimate |
With 36 GB unified memory
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| gemma-4-12B-it | Q8_0 | 12.7 GB | 27.4 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 36.4 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 77.4 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 36.4 estimate |
| Muse-Glimmer-30B | UD-Q5_K_M | 19.2 GB | 18.1 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 81.1 estimate |
| gemma-4-26B-A4B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 14.2 GB | 79.2 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 51.7 estimate |
With 64 GB unified memory
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| gemma-4-12B-it | Q8_0 | 12.7 GB | 27.4 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 36.4 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 77.4 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 36.4 estimate |
| Muse-Glimmer-30B | Q8_0 | 29.6 GB | 11.7 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 81.1 estimate |
| gemma-4-26B-A4B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 14.2 GB | 79.2 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 51.7 estimate |
With 128 GB unified memory
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| gemma-4-12B-it | Q8_0 | 12.7 GB | 27.4 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 36.4 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 77.4 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 36.4 estimate |
| Muse-Glimmer-30B | Q8_0 | 29.6 GB | 11.7 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 81.1 estimate |
| gemma-4-26B-A4B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 14.2 GB | 79.2 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 51.7 estimate |
Similar hardware
Apple M1 (7-core GPU) · Apple M1 (8-core GPU) · Apple M1 Pro (14-core GPU) · Apple M1 Pro (16-core GPU) · Apple M1 Max (24-core GPU) · Apple M1 Max (32-core GPU) · Apple M1 Ultra (48-core GPU) · Apple M1 Ultra (64-core GPU)