GraySoft
Model Comparison

mradermacher/gemma-4-31b-it-heretic-ara-i1-ggufvsprism-ml/bonsai-8b-gguf

Side-by-side comparison of mradermacher/gemma-4-31b-it-heretic-ara-i1-gguf and prism-ml/bonsai-8b-gguf: downloads, license, context length, tasks, and benchmarks.

mradermacher/gemma-4-31b-it-heretic-ara-i1-gguf

mradermacher · —

## About weighted/imatrix quants of https://huggingface.co/trohrbaugh/gemma-4-31b-it-heretic-ara ***For a convenient overview and download list, visit our model page for this model.*** static quants are available at https://huggingface.co/mradermacher/gemma-4-31b-it-heretic-ara-…

prism-ml/bonsai-8b-gguf

prism-ml · text-generation

End-to-end 1-bit language model for llama.cpp (CUDA, Metal, CPU) > **14.1x** smaller than FP16 | **6.2x** faster on RTX 4090 | **4-5x** lower energy/token

Side-by-side Specifications

mradermacher/gemma-4-31b-it-heretic-ara-i1-ggufprism-ml/bonsai-8b-gguf
Authormradermacherprism-ml
Pipeline Tasktext-generation
Librarytransformersllama.cpp
Downloads44,02883,309
Likes6618
LicenseUnknownUnknown
Context Length
Created2026-04-062026-03-18
Last Modified2026-04-072026-04-16
Tags
transformersggufhereticuncensoreddecensoredabliteratedaraenbase_model:trohrbaugh/gemma-4-31b-it-heretic-arabase_model:quantized:trohrbaugh/gemma-4-31b-it-heretic-ara
llama.cppgguf1-bitllama-cppcudametalon-deviceprismmlbonsaitext-generation

View full details: mradermacher/gemma-4-31b-it-heretic-ara-i1-gguf · prism-ml/bonsai-8b-gguf