GPUs & Macs › MediaTek Dimensity 9400
MediaTek Dimensity 9400: local LLM speed
Llama 2 7B (Q4_0) generates 8.9 tokens/s on the MediaTek Dimensity 9400 (Vulkan), and processes prompts at 38 tokens/s.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 2 7B | 8.9 | 38 | Q4_0 | Vulkan FA off | tg128 | llama.cpp #10879 details |
What fits on the MediaTek Dimensity 9400
Estimates for this chip are scaled from a measured run: Llama 2 7B Q4_0 at 8.9 tok/s (Vulkan, source). Fit assumes the whole model in GPU memory with ~8K context.
Memory size unknown for this device.