What Models?

Pick the right model for your GPU — in seconds.

Arete 27B VRAM requirements: which GPUs can run it locally?

27.36B params 256K max context VisionTool useReasoning

The smallest GPU memory that runs Arete 27B well is 18 GB (Q4_K_M, e.g. M3 Pro). 47 of 132 GPUs run it well.

VRAM needed per quantization

QuantizationWeights16K context32K contextFull 256K
Q4_K_M16.0 GB17.0 GB18.0 GB31.6 GB
Q8_026.7 GB27.7 GB28.7 GB42.3 GB

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

Run it locally

Each command downloads the GGUF from Hugging Face, then Ollama starts a chat and llama.cpp starts an OpenAI-compatible server with a web UI on port 8080. Both fetch the vision projector (mmproj) automatically.

Ollama ollama run hf.co/bartowski/NousResearch_Arete-27B-GGUF:Q4_K_M
llama.cpp llama-server -hf bartowski/NousResearch_Arete-27B-GGUF:Q4_K_M
Ollama ollama run hf.co/bartowski/NousResearch_Arete-27B-GGUF:Q8_0
llama.cpp llama-server -hf bartowski/NousResearch_Arete-27B-GGUF:Q8_0

Cheapest GPUs that can run Arete 27B

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 Arete 27B

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

  • M3 Pro 18 GB Q4_K_M 32K context · ~8 tok/s
  • M2 24 GB Q4_K_M 130K context · ~5 tok/s
  • M3 24 GB Q4_K_M 130K context · ~5 tok/s
  • M4 24 GB Q4_K_M 130K context · ~7 tok/s
  • M4 Pro 24 GB Q4_K_M 130K context · ~15 tok/s
  • M5 24 GB Q4_K_M 130K context · ~8 tok/s
  • M5 Pro 24 GB Q4_K_M 130K context · ~17 tok/s
  • M6 24 GB Q4_K_M 130K context · ~9 tok/s
  • M1 Max 32 GB Q8_0 86K context · ~13 tok/s
  • M1 Pro 32 GB Q8_0 86K context · ~7 tok/s
  • M2 Max 32 GB Q8_0 86K context · ~13 tok/s
  • M2 Pro 32 GB Q8_0 86K context · ~7 tok/s
  • M3 Max 36 GB Q8_0 152K context · ~10 tok/s
  • M4 Max 36 GB Q8_0 152K context · ~14 tok/s
  • M5 Max 36 GB Q8_0 152K context · ~16 tok/s
  • M1 Ultra 64 GB Q8_0 256K context · ~19 tok/s
  • M2 Ultra 64 GB Q8_0 256K context · ~19 tok/s
  • M3 Ultra 96 GB Q8_0 256K context · ~19 tok/s
  • M5 Ultra 96 GB Q8_0 256K context · ~28 tok/s
Check your exact setup in the full tool →