GPUs & Macs › Apple M4 (8-core GPU)
Apple M4 (8-core GPU): local LLM speed
3 measured results for the Apple M4 (8-core GPU).
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 3.2 1B Instruct | 65.9 | 898 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Llama 3.1 8B Instruct | 15.3 | 134 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Qwen2.5 14B Instruct | 7.3 | 61 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
What fits on the Apple M4 (8-core GPU)
Estimates for this chip are scaled from a measured run: Llama 3.1 8B Instruct Q4_K_M at 15.3 tok/s (llamafile, 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 |
|---|---|---|---|
| Qwen3.8-27B | UD-IQ2_XXS | 7.3 GB | 10.3 estimate |
| gemma-4-12B-it | Q6_K | 9.8 GB | 7.7 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 7.9 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 16.8 estimate |
| Qwen3.6-35B-A3B | UD-IQ1_M | 10.0 GB | 43.6 estimate |
| gemma-4-26B-A4B-it | UD-IQ2_M | 10.0 GB | 24.4 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 7.9 estimate |
| gemma-4-E4B-it | Q8_0 | 8.2 GB | 9.2 estimate |
With 36 GB unified memory
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| Qwen3.8-27B | UD-Q6_K | 22.0 GB | 3.4 estimate |
| gemma-4-12B-it | Q8_0 | 12.7 GB | 5.9 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 7.9 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 16.8 estimate |
| Qwen3.6-35B-A3B | UD-Q5_K_S | 24.9 GB | 17.6 estimate |
| gemma-4-26B-A4B-it | UD-Q6_K | 23.2 GB | 10.5 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 7.9 estimate |
| gemma-4-E4B-it | Q8_0 | 8.2 GB | 9.2 estimate |
With 64 GB unified memory
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| Qwen3.8-27B | Q8_0 | 29.0 GB | 2.6 estimate |
| gemma-4-12B-it | Q8_0 | 12.7 GB | 5.9 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 7.9 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 16.8 estimate |
| Qwen3.6-35B-A3B | Q8_0 | 36.9 GB | 11.9 estimate |
| gemma-4-26B-A4B-it | Q8_0 | 26.9 GB | 9.1 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 7.9 estimate |
| gemma-4-E4B-it | Q8_0 | 8.2 GB | 9.2 estimate |
With 128 GB unified memory
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| Qwen3.8-27B | Q8_0 | 29.0 GB | 2.6 estimate |
| gemma-4-12B-it | Q8_0 | 12.7 GB | 5.9 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 7.9 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 16.8 estimate |
| Qwen3.6-35B-A3B | Q8_0 | 36.9 GB | 11.9 estimate |
| Qwen3.8-Flash-Next | UD-IQ3_XXS | 82.0 GB | 0.9 estimate |
| gemma-4-26B-A4B-it | Q8_0 | 26.9 GB | 9.1 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 7.9 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)