GMKtec EVO-X2 Ryzen AI Max+ 395 128GB vs Mac Studio M4 Max 128GB
For running models locally, priced in the UK. At 64k context the first runs 40 models fully in memory, the second 40.
Specifications
| Spec | GMKtec EVO-X2 Ryzen AI Max+ 395 128GB | Mac Studio M4 Max 128GB |
|---|---|---|
| Chip | Ryzen AI Max+ 395 (Radeon 8060S) | Apple M4 Max (40-core GPU) |
| Memory bus | 256-bit | 512-bit |
| Memory bandwidth | 256 GB/s | 546 GB/s |
| FP16 compute | 59.4 TFLOPS | 16.16 TFLOPS |
| Software | ROCm | Metal |
| Unified memory | 128 GB | 128 GB |
| GPU can use by default | 96 GB | 96 GB |
| GPU can use, tuned | 120 GB | 120 GB |
| Operating system | windows | macos |
| Processor | Ryzen AI Max+ 395 | M4 Max 16-core |
| Storage | 2000 GB | 1000 GB |
What each runs
Estimated decode speed with the most precise quant that fits each one est.
| Model | GMKtec EVO-X2 Ryzen AI Max+ 395 128GB | Mac Studio M4 Max 128GB |
|---|---|---|
| Qwen 3.8 Flash Next | 3tok/sUD-Q3_K_XL | 7tok/sUD-Q3_K_XL |
| MiniMax M2.7 | 4tok/sUD-Q2_K_XL | 11tok/sUD-Q2_K_XL |
| Qwen 3.8 27B | 4tok/sQ8_0 | 11tok/sQ8_0 |
| Qwen 3.5 122B-A10B | 9tok/sQ5_K_M | 24tok/sQ5_K_M |
| Qwen 3.6 27B | 4tok/sQ8_0 | 11tok/sQ8_0 |
| gpt-oss 120B | 15tok/sMXFP4 | 38tok/sMXFP4 |
| Qwen3 Coder Next | 15tok/sQ8_0 | 37tok/sQ8_0 |
| Nemotron 3 Super 120B-A12B | 10tok/sUD-Q4_K_M | 24tok/sUD-Q4_K_M |
| Gemma 4 31B | 4tok/sQ8_0 | 9tok/sQ8_0 |
| Mistral Small 4 119B | 11tok/sUD-Q6_K | 29tok/sUD-Q6_K |
| Qwen 3.6 35B-A3B | 19tok/sQ8_0 | 48tok/sQ8_0 |
| Qwen 3.5 27B | 4tok/sQ8_0 | 11tok/sQ8_0 |
| Gemma 4 26B-A4B | 15tok/sQ8_0 | 37tok/sQ8_0 |
| GLM 4.5 Air | 3tok/sQ5_K_M | 8tok/sQ5_K_M |
| Qwen 3.5 35B-A3B | 19tok/sQ8_0 | 48tok/sQ8_0 |
| Qwen3 Next 80B-A3B | 18tok/sQ8_0 | 44tok/sQ8_0 |
| GLM 4.7 Flash | 10tok/sQ8_0 | 24tok/sQ8_0 |
| Devstral Small 2 24B | 4tok/sQ8_0 | 10tok/sQ8_0 |
| Nemotron 3.5 Lightning 30B-A3B | 22tok/sUD-Q6_K_XL | 56tok/sUD-Q6_K_XL |
| Kimi Linear 48B-A3B | 23tok/sQ8_0 | 57tok/sQ8_0 |
| gpt-oss 20B | 21tok/sMXFP4 | 53tok/sMXFP4 |
| Qwen3 Coder 30B-A3B | 9tok/sQ8_0 | 22tok/sQ8_0 |
| Qwen3 32B | 3tok/sQ8_0 | 7tok/sQ8_0 |
| Gemma 4 12B | 10tok/sQ8_0 | 25tok/sQ8_0 |
| Qwen 3.5 9B | 12tok/sQ8_0 | 30tok/sQ8_0 |
| Gemma 3 27B | 4tok/sQ8_0 | 10tok/sQ8_0 |
| Mistral Small 3.2 24B | 4tok/sQ8_0 | 10tok/sQ8_0 |
| DeepSeek R1 Distill Qwen 32B | 3tok/sQ8_0 | 7tok/sQ8_0 |
| Llama 3.3 70B | 2tok/sQ8_0 | 4tok/sQ8_0 |
| Ministral 3 14B | 6tok/sQ8_0 | 14tok/sQ8_0 |
| Qwen3 14B | 5tok/sQ8_0 | 13tok/sQ8_0 |
| DeepSeek R1 0528 Qwen3 8B | 8tok/sQ8_0 | 19tok/sQ8_0 |
| Qwen3 8B | 8tok/sQ8_0 | 19tok/sQ8_0 |
| Phi-4 14B | 5tok/sQ8_0 | 12tok/sQ8_0 |
| Qwen 3.5 4B | 21tok/sQ8_0 | 53tok/sQ8_0 |
| Ministral 3 8B | 8tok/sQ8_0 | 20tok/sQ8_0 |
| Granite 4.2 8B | 7tok/sQ8_0 | 18tok/sQ8_0 |
| Gemma 4 E4B | 14tok/sQ8_0 | 35tok/sQ8_0 |
| LFM2.5 8B-A1B | 20tok/sQ8_0 | 50tok/sQ8_0 |
| Llama 3.1 8B | 9tok/sQ6_K | 23tok/sQ6_K |
| DeepSeek V4 Flash | won't run | won't run |
| GLM 5.3 Flash | won't run | won't run |