What Models?

Pick the right model for your GPU — in seconds.

What GPU do you need to run Llama 3.2 1B?

1.24B params 128K max context Tool use

The smallest GPU memory that runs Llama 3.2 1B well is 2 GB (Q8_0, e.g. T400). 132 of 132 GPUs run it well.

MMLU 49.3

VRAM needed per quantization

QuantizationWeights16K context32K contextFull 128K
Q4_K_M0.8 GB1.3 GB1.8 GB4.8 GB
Q8_01.3 GB1.8 GB2.3 GB5.3 GB

Weights plus fp16 KV cache. Leave ~1 GB headroom for the inference engine.

Cheapest GPUs that can run Llama 3.2 1B

Lowest launch MSRP among GPUs that run its smallest quantization (Q4_K_M) well, with at least 16K context.

As an Amazon Associate, I earn from qualifying purchases.

GPUs that can run Llama 3.2 1B

Best quantization that fits on each GPU, with the context and speed it gets. For chips with several memory sizes, the smallest that works is shown.

NVIDIA

AMD

Intel

Apple

  • M1 8 GB Q8_0 128K context · ~29 tok/s
  • M2 8 GB Q8_0 128K context · ~43 tok/s
  • M3 8 GB Q8_0 128K context · ~43 tok/s
  • M1 Pro 16 GB Q8_0 128K context · ~86 tok/s
  • M2 Pro 16 GB Q8_0 128K context · ~86 tok/s
  • M4 16 GB Q8_0 128K context · ~52 tok/s
  • M5 16 GB Q8_0 128K context · ~66 tok/s
  • M6 16 GB Q8_0 128K context · ~66 tok/s
  • M3 Pro 18 GB Q8_0 128K context · ~65 tok/s
  • M4 Pro 24 GB Q8_0 128K context · ~118 tok/s
  • M5 Pro 24 GB Q8_0 128K context · ~132 tok/s
  • M1 Max 32 GB Q8_0 128K context · ~172 tok/s
  • M2 Max 32 GB Q8_0 128K context · ~172 tok/s
  • M3 Max 36 GB Q8_0 128K context · ~129 tok/s
  • M4 Max 36 GB Q8_0 128K context · ~177 tok/s
  • M5 Max 36 GB Q8_0 128K context · ~198 tok/s
  • M1 Ultra 64 GB Q8_0 128K context · ~345 tok/s
  • M2 Ultra 64 GB Q8_0 128K context · ~345 tok/s
  • M3 Ultra 96 GB Q8_0 128K context · ~353 tok/s
  • M5 Ultra 96 GB Q8_0 128K context · ~517 tok/s
Check your exact setup in the full tool →