GGUF models by BeaverAI
381 GGUF models quantized by BeaverAI on Hugging Face. Sizes run from 1.72B to 141B parameters. The longest trained context is 1000K tokens.
About BeaverAI on Hugging Face
BeaverAI has published 381 GGUF models that ggufmodels.com tracks, across families such as Cydonia, Gemma and Behemoth. The most downloaded are Spectre 37B v1a, Spectre 37B v1a GGUF Dirty and Betta 30B v1d.
The quants BeaverAI publishes most often are Q4_K_M (348), Q5_K_M (332), Q8_0 (324) and Q6_K (324). Their Q4_K_M files range from 1.1 to 85.6 GB. Parameter counts run from 1.72B to 141B, and the longest trained context window is 1000K tokens.
Frequently asked questions
What GGUF models has BeaverAI published?
381 models are listed here, mainly from the Cydonia, Gemma and Behemoth families. The most downloaded is Spectre 37B v1a.
Which quants does BeaverAI publish?
The most common are Q4_K_M (348 models), Q5_K_M (332 models), Q8_0 (324 models) and Q6_K (324 models).