F32 GGUF models
Unquantized 32-bit floats. Rarely needed for running a model. 2087 models have a F32 file. Sizes run from 1M to 310B parameters. The longest trained context is 10240K tokens.
What is F32?
Unquantized 32-bit floats. Rarely needed for running a model. On average it stores about 32 bits per weight, so a F32 file is roughly 200% of the size of the same model in 16-bit. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.
2,087 models have a F32 file on ggufmodels.com, and those files range from 0.0 to 943.0 GB. The most downloaded are YuE2 audio tokenizer, the v9 head of Mothersuperior, on MERT-v2-FullSong, Parakeet Unified EN 0.6B and nemotron 3.5 asr streaming 0.6b. For a longer explanation of every quant type, read GGUF quantization explained.
Frequently asked questions
What is F32 in GGUF?
Unquantized 32-bit floats. Rarely needed for running a model. It averages about 32 bits per weight.
Should I use F32?
The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.
How big is a F32 file?
F32 files on this site range from 0.0 to 943.0 GB, depending on the model. As a rule of thumb it is about 200% of the 16-bit size.