GPUs & Macs › AMD Ryzen AI Max+ 395 (Radeon 8060S)
AMD Ryzen AI Max+ 395 (Radeon 8060S): local LLM speed
Llama 2 7B (Q4_0) generates 54.4 tokens/s on the AMD Ryzen AI Max+ 395 (Radeon 8060S) (Vulkan), and processes prompts at 1,704 tokens/s.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 2 7B | 54.4 | 1,704 | Q4_0 | Vulkan FA on | tg128 | llama.cpp #10879 details |
| Llama 2 7B | 51.8 | 1,487 | Q4_0 | Vulkan FA off | tg128 | llama.cpp #10879 details |
| Llama 2 7B | 49.4 | 685 | Q4_0 | ROCm FA on | tg128 · n=2 | llama.cpp #15021 details |
| Llama 2 7B | 49.0 | 631 | Q4_0 | ROCm FA off | tg128 · n=2 | llama.cpp #15021 details |
What fits on the AMD Ryzen AI Max+ 395 (Radeon 8060S)
Estimates for this chip are scaled from a measured run: Llama 2 7B Q4_0 at 54.4 tok/s (Vulkan, source). Fit assumes the whole model in GPU memory with ~8K context.
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| Qwen3.8-27B | Q8_0 | 29.0 GB | 7.2 estimate |
| gemma-4-12B-it | Q8_0 | 12.7 GB | 16.4 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 21.8 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 46.4 estimate |
| Qwen3.6-35B-A3B | Q8_0 | 36.9 GB | 32.9 estimate |
| Qwen3.8-Flash-Next | UD-IQ3_XXS | 82.0 GB | 2.5 estimate |
| gemma-4-26B-A4B-it | Q8_0 | 26.9 GB | 25.2 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 21.8 estimate |
| gemma-4-E4B-it | Q8_0 | 8.2 GB | 25.4 estimate |
| Muse-Glimmer-30B | Q8_0 | 29.6 GB | 7.0 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 48.6 estimate |
| gemma-4-31B-it | Q8_0 | 32.6 GB | 6.4 estimate |
| gemma-4-26B-A4B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 14.2 GB | 47.4 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 31.0 estimate |
Similar hardware
AMD BC-250 · AMD Radeon RX 6800M · AMD Radeon RX 6600M · AMD Radeon RX 7600M XT · AMD Ryzen AI 300 Series · AMD Ryzen 6000 Series · AMD Ryzen 7000 Series · AMD Ryzen 8000 Series