GraySoft
Download guIDE →
Model Comparison

mudler/gemma-4-26b-a4b-it-heretic-apex-ggufvsunsloth/gemma-4-31b-it-gguf

Side-by-side comparison of mudler/gemma-4-26b-a4b-it-heretic-apex-gguf and unsloth/gemma-4-31b-it-gguf: downloads, license, context length, tasks, and benchmarks.

mudler/gemma-4-26b-a4b-it-heretic-apex-gguf

mudler · —

**APEX (Adaptive Precision for EXpert Models)** quantizations of gemma-4-26B-A4B-it-heretic — an **abliterated (uncensored)** version of Gemma 4, created with the Heretic tool (v1.2.0) using Arbitrary-Rank Ablation (ARA) on layers 10-30 to reduce refusals while preserving capabi…

unsloth/gemma-4-31b-it-gguf

unsloth · image-text-to-text

processor = AutoProcessor.from_pretrained(MODEL_ID) model = AutoModelForCausalLM.from_pretrained( MODEL_ID, dtype="auto", device_map="auto" ) `` Once the model is loaded, you can start generating output: `python # Prompt messages = [ {"role": "system", "content": "You are a help…

Side-by-side Specifications

mudler/gemma-4-26b-a4b-it-heretic-apex-ggufunsloth/gemma-4-31b-it-gguf
Authormudlerunsloth
Pipeline Task—image-text-to-text
Library——
Downloads18,8491,362,626
Likes22325
LicenseUnknownUnknown
Context Length——
Created2026-04-052026-04-01
Last Modified2026-04-052026-04-11
Tags
ggufquantizedapexmoemixture-of-expertsgemma4vlmvisionbase_model:coder3101/gemma-4-26B-A4B-it-hereticbase_model:quantized:coder3101/gemma-4-26B-A4B-it-heretic
ggufgemma4unslothgemmagoogleimage-text-to-textbase_model:google/gemma-4-31B-itbase_model:quantized:google/gemma-4-31B-itlicense:apache-2.0endpoints_compatible

View full details: mudler/gemma-4-26b-a4b-it-heretic-apex-gguf · unsloth/gemma-4-31b-it-gguf