Model Anatomy See what's inside open-weight AI models

Can I run clef on a Mac with 64 GB?

Yes, with a GGUF version

The published weights are too big, but the Q8_0 file from bartowski/Cloudflare_clef-GGUF (28.8 GB) fits, with room for ≈ 256K tokens of context.

Full analysis of clef Everything that runs on Mac with 64 GB Get the Q8_0 file on Hugging Face
ollama run hf.co/bartowski/Cloudflare_clef-GGUF:Q8_0

Files and context on Mac with 64 GB

Each file of clef (and of GGUF versions the site has analyzed), and how much context fits next to it, with the context cache in 16-bit or 8-bit.

FileSizeOn Mac with 64 GB Context (16-bit)Context (8-bit)
As published54.7 GBDoesn't fit——
BF16bartowski/Cloudflare_clef-GGUF54.1 GBDoesn't fit——
Q8_0bartowski/Cloudflare_clef-GGUF28.8 GBFits256K256K
Q6_K_Lbartowski/Cloudflare_clef-GGUF24.0 GBFits256K256K
Q6_Kbartowski/Cloudflare_clef-GGUF23.3 GBFits256K256K
Q5_K_Mbartowski/Cloudflare_clef-GGUF20.6 GBFits256K256K
Q5_K_Sbartowski/Cloudflare_clef-GGUF19.6 GBFits256K256K
Q4_K_Lbartowski/Cloudflare_clef-GGUF18.6 GBFits256K256K
Q4_1bartowski/Cloudflare_clef-GGUF17.7 GBFits256K256K
Q4_K_Mbartowski/Cloudflare_clef-GGUF17.7 GBFits256K256K
Q4_K_Sbartowski/Cloudflare_clef-GGUF16.6 GBFits256K256K
IQ4_NLbartowski/Cloudflare_clef-GGUF16.3 GBFits256K256K
Q4_0bartowski/Cloudflare_clef-GGUF16.2 GBFits256K256K
IQ4_XSbartowski/Cloudflare_clef-GGUF15.6 GBFits256K256K
Q3_K_Lbartowski/Cloudflare_clef-GGUF15.2 GBFits256K256K
Q3_K_Mbartowski/Cloudflare_clef-GGUF14.5 GBFits256K256K
IQ3_Mbartowski/Cloudflare_clef-GGUF13.8 GBFits256K256K
Q3_K_Sbartowski/Cloudflare_clef-GGUF13.6 GBFits256K256K
Q2_Kbartowski/Cloudflare_clef-GGUF11.7 GBFits256K256K
BF16ggml-org/Clef-GGUF54.1 GBDoesn't fit——
Q8_0ggml-org/Clef-GGUF28.7 GBFits256K256K
Q4_K_Mggml-org/Clef-GGUF19.2 GBFits256K256K

clef on other hardware

Popular models that run on Mac with 64 GB

Assumes 75% of the Mac's memory is usable (macOS's default limit for the GPU; it can be raised) and at least 4K tokens of context for one conversation, with the context cache in 16-bit unless noted. Fits leaves at least 20% of that memory for context; Tight leaves less.