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

Can I run NeoHorse-1-4B on an RTX 4060?

Yes, with a GGUF version

The published weights are too big, but the Q8_0 file from TokenRhythm/NeoHorse-1-4B-GGUF (4.48 GB) fits, with room for ≈ 79K tokens of context.

Full analysis of NeoHorse-1-4B Everything that runs on RTX 4060 Get the Q8_0 file on Hugging Face
ollama run hf.co/TokenRhythm/NeoHorse-1-4B-GGUF:Q8_0

Files and context on RTX 4060

Each file of NeoHorse-1-4B (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 RTX 4060 Context (16-bit)Context (8-bit)
As published8.41 GBDoesn't fit——
BF16TokenRhythm/NeoHorse-1-4B-GGUF8.42 GBDoesn't fit——
F16TokenRhythm/NeoHorse-1-4B-GGUF8.42 GBDoesn't fit——
Q8_0TokenRhythm/NeoHorse-1-4B-GGUF4.48 GBFits79K159K
Q5_K_MTokenRhythm/NeoHorse-1-4B-GGUF3.07 GBFits121K243K
Q4_K_MTokenRhythm/NeoHorse-1-4B-GGUF2.71 GBFits132K256K

NeoHorse-1-4B on other hardware

Popular models that run on RTX 4060

Assumes 90% of the card's memory is usable 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.