Per-model VRAM calculators

Exact on-disk quantization sizes synced from Hugging Face, with estimated total VRAM per quantization and context length. See the generic calculator on /hardware if your model isn't listed here.

DeepSeek-R1-Distill-Llama-8B-GGUF

24 quantizations synced

DeepSeek-R1-Distill-Qwen-14B-GGUF

26 quantizations synced

DeepSeek-R1-Distill-Qwen-32B-GGUF

26 quantizations synced

Devstral-Small-2507-GGUF

26 quantizations synced

gemma-2-27b-it-GGUF

25 quantizations synced

gemma-2-9b-it-GGUF

29 quantizations synced

GLM-4.7-Flash-GGUF

28 quantizations synced

google_gemma-4-26B-A4B-it-GGUF

27 quantizations synced

google_gemma-4-31B-it-GGUF

28 quantizations synced

gpt-oss-20b-GGUF

16 quantizations synced

Meta-Llama-3-8B-Instruct-GGUF

14 quantizations synced

Meta-Llama-3.1-70B-Instruct-GGUF

23 quantizations synced

Meta-Llama-3.1-8B-Instruct-GGUF

24 quantizations synced

Mistral-7B-Instruct-v0.3-GGUF

23 quantizations synced

Mixtral-8x7B-Instruct-v0.1-GGUF

8 quantizations synced

Phi-3-mini-4k-instruct-GGUF

5 quantizations synced

Qwen2.5-32B-Instruct-GGUF

9 quantizations synced

Qwen2.5-7B-Instruct-GGUF

9 quantizations synced

Qwen3-Coder-30B-A3B-Instruct-GGUF

27 quantizations synced

Qwen3.6-27B-GGUF

22 quantizations synced

Qwen3.8-27B-GGUF

28 quantizations synced