GraySoft
Download guIDE →
Model Comparison

noctrex/gemma-4-26b-a4b-it-mxfp4_moe-ggufvsprism-ml/bonsai-8b-gguf

Side-by-side comparison of noctrex/gemma-4-26b-a4b-it-mxfp4_moe-gguf and prism-ml/bonsai-8b-gguf: downloads, license, context length, tasks, and benchmarks.

noctrex/gemma-4-26b-a4b-it-mxfp4_moe-gguf

noctrex · image-text-to-text

These are **MXFP4** quantizations of the model gemma-4-26B-A4B-it

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

noctrex/gemma-4-26b-a4b-it-mxfp4_moe-ggufprism-ml/bonsai-8b-gguf
Authornoctrexprism-ml
Pipeline Taskimage-text-to-texttext-generation
Library—llama.cpp
Downloads19,21483,309
Likes10618
LicenseUnknownUnknown
Context Length——
Created2026-04-022026-03-18
Last Modified2026-04-122026-04-16
Tags
ggufimage-text-to-textbase_model:google/gemma-4-26B-A4B-itbase_model:quantized:google/gemma-4-26B-A4B-itendpoints_compatibleregion:usconversational
llama.cppgguf1-bitllama-cppcudametalon-deviceprismmlbonsaitext-generation

View full details: noctrex/gemma-4-26b-a4b-it-mxfp4_moe-gguf · prism-ml/bonsai-8b-gguf