GPUs & Macs › Intel Core Ultra 300 Series
Intel Core Ultra 300 Series: local LLM speed
Llama 2 7B (Q4_0) generates 24.4 tokens/s on the Intel Core Ultra 300 Series (Vulkan), and processes prompts at 1,380 tokens/s.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 2 7B | 24.4 | 1,380 | Q4_0 | Vulkan FA off | tg128 | llama.cpp #10879 details |
| Llama 2 7B | 23.9 | 1,332 | Q4_0 | Vulkan FA on | tg128 | llama.cpp #10879 details |
What fits on the Intel Core Ultra 300 Series
Estimates for this chip are scaled from a measured run: Llama 2 7B Q4_0 at 24.4 tok/s (Vulkan, source). Fit assumes the whole model in GPU memory with ~8K context.
Memory size unknown for this device.
Similar hardware
Intel Arc A770M · Intel Core Ultra 200 Series · Intel Core 1100 Series · Intel Core Ultra 100 Series · Intel Core 1000 Series · Intel Core 8000 Series · Intel N150