Framework Desktop Ryzen AI Max+ 395 128GB vs GMKtec EVO-X2 Ryzen AI Max+ 395 128GB
For running models locally, priced in India. At 8k context the first runs 41 models fully in memory, the second 41.
In short
Both have 128 GB of memory for models; the Framework Desktop Ryzen AI Max+ 395 128GB has 256 GB/s of bandwidth and the GMKtec EVO-X2 Ryzen AI Max+ 395 128GB 256 GB/s. At 8k context both run the same 41 models we track fully in memory. On the models both run fully in memory, the Framework Desktop Ryzen AI Max+ 395 128GB and the GMKtec EVO-X2 Ryzen AI Max+ 395 128GB are about as fast as each other (estimated).
Updated . Speeds are estimates; how they are worked out.
Specifications
| Spec | Framework Desktop Ryzen AI Max+ 395 128GB | GMKtec EVO-X2 Ryzen AI Max+ 395 128GB |
|---|---|---|
| Chip | Ryzen AI Max+ 395 (Radeon 8060S) | Ryzen AI Max+ 395 (Radeon 8060S) |
| Memory bus | 256-bit | 256-bit |
| Memory bandwidth | 256 GB/s | 256 GB/s |
| FP16 compute | 59.4 TFLOPS | 59.4 TFLOPS |
| Software | ROCm | ROCm |
| 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 | linux | windows |
| Processor | Ryzen AI Max+ 395 | Ryzen AI Max+ 395 |
| Storage | 1000 GB | 2000 GB |
What each runs
Estimated decode speed with the most precise quant that fits each one est.
| Model | Framework Desktop Ryzen AI Max+ 395 128GB | GMKtec EVO-X2 Ryzen AI Max+ 395 128GB |
|---|---|---|
| DeepSeek V4 Flash | 11tok/sUD-Q2_K_XL | 11tok/sUD-Q2_K_XL |
| Qwen 3.8 Flash Next | 3tok/sUD-Q3_K_XL | 3tok/sUD-Q3_K_XL |
| MiniMax M2.7 | 15tok/sUD-Q2_K_XL | 15tok/sUD-Q2_K_XL |
| Qwen 3.8 27B | 5tok/sQ8_0 | 5tok/sQ8_0 |
| Qwen 3.5 122B-A10B | 10tok/sQ6_K | 10tok/sQ6_K |
| Qwen 3.6 27B | 5tok/sQ8_0 | 5tok/sQ8_0 |
| gpt-oss 120B | 25tok/sMXFP4 | 25tok/sMXFP4 |
| Qwen3 Coder Next | 19tok/sQ8_0 | 19tok/sQ8_0 |
| Nemotron 3 Super 120B-A12B | 10tok/sUD-Q4_K_M | 10tok/sUD-Q4_K_M |
| Gemma 4 31B | 4tok/sQ8_0 | 4tok/sQ8_0 |
| Mistral Small 4 119B | 14tok/sUD-Q6_K | 14tok/sUD-Q6_K |
| Qwen 3.6 35B-A3B | 26tok/sQ8_0 | 26tok/sQ8_0 |
| Qwen 3.5 27B | 5tok/sQ8_0 | 5tok/sQ8_0 |
| Gemma 4 26B-A4B | 19tok/sQ8_0 | 19tok/sQ8_0 |
| GLM 4.5 Air | 5tok/sQ6_K | 5tok/sQ6_K |
| Qwen 3.5 35B-A3B | 26tok/sQ8_0 | 26tok/sQ8_0 |
| Qwen3 Next 80B-A3B | 25tok/sQ8_0 | 25tok/sQ8_0 |
| GLM 4.7 Flash | 15tok/sQ8_0 | 15tok/sQ8_0 |
| Devstral Small 2 24B | 5tok/sQ8_0 | 5tok/sQ8_0 |
| Nemotron 3.5 Lightning 30B-A3B | 25tok/sUD-Q6_K_XL | 25tok/sUD-Q6_K_XL |
| gpt-oss 20B | 32tok/sMXFP4 | 32tok/sMXFP4 |
| Kimi Linear 48B-A3B | 26tok/sQ8_0 | 26tok/sQ8_0 |
| Qwen3 Coder 30B-A3B | 21tok/sQ8_0 | 21tok/sQ8_0 |
| Qwen3 32B | 4tok/sQ8_0 | 4tok/sQ8_0 |
| Gemma 4 12B | 11tok/sQ8_0 | 11tok/sQ8_0 |
| Qwen 3.5 9B | 14tok/sQ8_0 | 14tok/sQ8_0 |
| Mistral Small 3.2 24B | 5tok/sQ8_0 | 5tok/sQ8_0 |
| Gemma 3 27B | 5tok/sQ8_0 | 5tok/sQ8_0 |
| DeepSeek R1 Distill Qwen 32B | 4tok/sQ8_0 | 4tok/sQ8_0 |
| Llama 3.3 70B | 2tok/sQ8_0 | 2tok/sQ8_0 |
| Ministral 3 14B | 9tok/sQ8_0 | 9tok/sQ8_0 |
| Qwen3 14B | 8tok/sQ8_0 | 8tok/sQ8_0 |
| DeepSeek R1 0528 Qwen3 8B | 14tok/sQ8_0 | 14tok/sQ8_0 |
| Qwen3 8B | 14tok/sQ8_0 | 14tok/sQ8_0 |
| Phi-4 14B | 8tok/sQ8_0 | 8tok/sQ8_0 |
| Qwen 3.5 4B | 29tok/sQ8_0 | 29tok/sQ8_0 |
| Ministral 3 8B | 14tok/sQ8_0 | 14tok/sQ8_0 |
| Granite 4.2 8B | 13tok/sQ8_0 | 13tok/sQ8_0 |
| Gemma 4 E4B | 17tok/sQ8_0 | 17tok/sQ8_0 |
| LFM2.5 8B-A1B | 58tok/sQ8_0 | 58tok/sQ8_0 |
| Llama 3.1 8B | 18tok/sQ6_K | 18tok/sQ6_K |
| GLM 5.3 Flash | won't run | won't run |
Questions
- Which can run bigger models, the Framework Desktop Ryzen AI Max+ 395 128GB or the GMKtec EVO-X2 Ryzen AI Max+ 395 128GB?
- At 8k context both run the same 41 models we track fully in memory.
- Which is faster for local AI, the Framework Desktop Ryzen AI Max+ 395 128GB or the GMKtec EVO-X2 Ryzen AI Max+ 395 128GB?
- On the models both run fully in memory, the Framework Desktop Ryzen AI Max+ 395 128GB and the GMKtec EVO-X2 Ryzen AI Max+ 395 128GB are about as fast as each other (estimated).