This page provides GGUF quantized versions of DiffusionGemma 26B A4B-it, a multimodal model from Google DeepMind based on the Gemma 4 architecture. The model employs discrete text diffusion through block-autoregressive multi-canvas sampling to achieve significantly faster decoding speeds than standard autoregressive models. It is capable of processing interleaved inputs consisting of text, images with variable resolutions and aspect ratios, and video content for generating textual outputs.
Key topics:
- Mixture-of-Experts architecture with 3.8 billion active parameters.
- High-speed generation through parallel denoising of token blocks.
- Multimodal input support including image and video understanding.
- Extensive context window capability up to 256K tokens.
- Integrated reasoning modes for step-by-step thought processes.
Key topics:
- Mixture-of-Experts architecture with 3.8 billion active parameters.
- High-speed generation through parallel denoising of token blocks.
- Multimodal input support including image and video understanding.
- Extensive context window capability up to 256K tokens.
- Integrated reasoning modes for step-by-step thought processes.
This guide provides instructions for running Alibaba's Qwen3.6 multimodal hybrid-thinking models locally using Unsloth tools. It covers the 27B and 35B-A3B variants, which support a 256K context window across 201 languages and excel in agentic coding, vision, and chat tasks. The article details hardware requirements for various quantization levels and explains how to leverage Multi Token Prediction (MTP) for significantly faster inference.
Key topics:
- Hardware memory requirements for quantized models
- Faster generation via Multi Token Prediction (MTP)
- Integration with Unsloth Studio, llama.cpp, and MLX
- Preserved thinking mode configurations
Key topics:
- Hardware memory requirements for quantized models
- Faster generation via Multi Token Prediction (MTP)
- Integration with Unsloth Studio, llama.cpp, and MLX
- Preserved thinking mode configurations
This repository provides the GGUF quantized weights for Qwen3.6-27B, a flagship-level coding model designed for stability and real-world utility. The model features significant upgrades in agentic coding capabilities, allowing it to handle frontend workflows and repository-level reasoning with high precision. It also introduces thinking preservation, which enables the model to retain reasoning context from historical messages to improve iterative development.
Key technical highlights:
* Native context length of 262,144 tokens, extensible up to 1,010,000 via RoPE scaling (YaRN).
* Enhanced tool-calling capabilities for complex agentic tasks.
* Support for multimodal inputs including images and video.
* Optimized for various inference frameworks like SGLang, vLLM, and KTransformers.
Key technical highlights:
* Native context length of 262,144 tokens, extensible up to 1,010,000 via RoPE scaling (YaRN).
* Enhanced tool-calling capabilities for complex agentic tasks.
* Support for multimodal inputs including images and video.
* Optimized for various inference frameworks like SGLang, vLLM, and KTransformers.
Unsloth AI presents performance benchmarks for Qwen3.6-35B-A3B GGUF quantizations, claiming state-of-the-art results in mean KL divergence across most model sizes. The discussion includes community analysis regarding SWE-bench Verified performance, where some users noted unexpected discrepancies between Qwen3.5 and Qwen3.6 quantization results during coding tasks.
Key points:
- Unsloth ranks first in 21 of 22 model sizes for mean KL divergence.
- Community debate over SWE-bench testing methodology and sample sizes.
- Reported performance variations between different quantization levels (Q4, Q5, Q6, Q8).
- Discussion on system prompt adherence and error rates in coding benchmarks.
Key points:
- Unsloth ranks first in 21 of 22 model sizes for mean KL divergence.
- Community debate over SWE-bench testing methodology and sample sizes.
- Reported performance variations between different quantization levels (Q4, Q5, Q6, Q8).
- Discussion on system prompt adherence and error rates in coding benchmarks.
This document details how to run Google's Gemma 4 models locally, including the E2B, E4B, 26B-A4B, and 31B variants. Gemma 4 is a family of open models supporting over 140 languages and up to 256K context, available in both dense and MoE configurations. The E2B and E4B models support image and audio input. These models can be run locally on your device and fine-tuned using Unsloth Studio. The document outlines hardware requirements, recommended settings, and best practices for prompting and multimodal use, including guidance on context length and thinking mode.
This Hugging Face page details the Gemma 4 31B-it model, an open-weights multimodal model created by Google DeepMind. Gemma 4 can process both text and image inputs, generating text outputs, with smaller models also supporting audio. It comes in various sizes (E2B, E4B, 26B A4B, and 31B) allowing for deployment on diverse hardware, from phones to servers.
The model boasts a context window of up to 256K tokens and supports over 140 languages. It utilizes dense and Mixture-of-Experts (MoE) architectures, excelling in tasks like text generation, coding, and reasoning. The page provides details on model data, training, ethics, usage, limitations, and best practices, along with code snippets for getting started with Transformers.
The model boasts a context window of up to 256K tokens and supports over 140 languages. It utilizes dense and Mixture-of-Experts (MoE) architectures, excelling in tasks like text generation, coding, and reasoning. The page provides details on model data, training, ethics, usage, limitations, and best practices, along with code snippets for getting started with Transformers.
This article details benchmarks for Unsloth Dynamic GGUFs of the Qwen3.5 model, including analysis of perplexity, KL divergence, and MXFP4. It covers performance across different bit widths and quant types, highlighting the impact of Imatrix and the limitations of certain quantization approaches. Full benchmark data is also provided.
This guide explains how to use tool calling with local LLMs, including examples with mathematical, story, Python code, and terminal functions, using llama.cpp, llama-server, and OpenAI endpoints.
This article details the performance of Unsloth Dynamic GGUFs on the Aider Polyglot benchmark, showcasing how it can quantize LLMs like DeepSeek-V3.1 to as low as 1-bit while outperforming models like GPT-4.5 and Claude-4-Opus. It also covers benchmark setup, comparisons to other quantization methods, and chat template bug fixes.
How to run Gemma 3 effectively with our GGUFs on llama.cpp, Ollama, Open WebUI and how to fine-tune with Unsloth! This page details running Gemma 3 on various platforms, including phones, and fine-tuning it using Unsloth, addressing potential issues with float16 precision and providing optimal configuration settings.
