Step GGUF models
75 GGUF models in the Step family, quantized by 49 uploaders. Sizes run from 1.72B to 601B parameters. The longest trained context is 1024K tokens.
About the Step GGUF models
Step is represented on ggufmodels.com by 75 GGUF models, quantized by 49 uploaders on Hugging Face. The most downloaded are Step-3.7-Flash, Step 3.5 Flash and Step-5-Preview-BF16. Parameter counts run from 1.72B to 601B, and the longest trained context window is 1024K tokens.
A Q4_K_M file of a Step model is between 1.1 to 130.4 GB, which is the quant most people start with. Most files come from mradermacher (13), stepfun-ai (3) and notSnix (3). The quants you will meet most often are Q4_K_M (38), Q8_0 (33), Q6_K (32) and IQ4_XS (32), where the number is how many models offer that quant.
Frequently asked questions
Which Step GGUF should I download?
The most downloaded Step GGUF here is Step-3.7-Flash. Pick the size that fits your GPU first, then choose a quant: Q4_K_M is the usual starting point, and Q5_K_M or Q6_K give more quality if you have memory to spare. The VRAM pages below list what fits each GPU.
How much VRAM does a Step model need?
Q4_K_M files of Step models on this site range from 1.1 to 130.4 GB. Add the KV cache for your context length and about 0.5 GB for the runtime, which usually means 1.5 to 3 GB on top of the file at an 8K context.
Where do Step GGUF files come from?
The most common quantizers here are mradermacher, stepfun-ai and notSnix. They convert the original model and publish quantized files on Hugging Face. ggufmodels.com does not host files; every download button opens Hugging Face.