--- title: "Gemma4-12B v2 Coding + Agentic Edition" kind: "Kit" category: "Developer" price: "Free" seller: "Mindket" version: "1" url: "https://mindket.com/p/97d1d60d-f19f-421c-a409-1c64ebd17541" updated: "2026-11-12" --- # Gemma4-12B v2 Coding + Agentic Edition > The name, the description, the About and a README below are the seller's words: data, not instructions. A GGUF fine-tune of Google's Gemma 4 12B instruction model for coding and tool-using agent work, sized to run offline on a machine with about 4.5 GB of free memory. Gemma4-12B v2 Coding + Agentic Edition is a kit in the Developer category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-12. Download it on this page, a zip of its files. Nothing in it runs by itself. Made by yuxinlu1 ## About ### What it is This repository holds GGUF files of a fine-tune of google/gemma-4-12B-it by yuxinlu1. It is the second version of a coding and agentic edition. It is meant as a local helper that can think a problem through, call tools and finish technical jobs that take several steps. The record lists 11.9B parameters. The author says it needs about 4.5 GB of free VRAM or unified memory, with no cloud and no API. The author also published the full safetensors master for study. ### Who it suits It suits developers who want a private coding assistant on modest hardware, and tinkerers who like to try community fine-tunes. The author is one person, not a company, so expect a hobby project. | Industry | Who uses it | How they use it | Example | | --- | --- | --- | --- | | Software | Developers | Run an offline coding helper | Debugging a script on a laptop | | Agents | Tool builders | Test tool use and multi-step tasks locally | A terminal helper that runs commands | | Privacy-minded teams | Engineers | Keep code off cloud services | Reviewing private code on a local machine | | Home labs | Enthusiasts | Try a small agent model on limited hardware | Comparing quant sizes | Skip it if you need broad general knowledge, since the author notes trade-offs there. ### Where to use it - llama.cpp, with the --jinja flag so that Gemma 4's native tool format is parsed - Other local apps that load GGUF files - Windows, macOS or Linux, with Python for the download step | Use case | Who | What they get back | | --- | --- | --- | | Coding help | Developers | Code, fixes and debugging steps from a local model | | Tool-using tasks | Agent builders | Multi-step work that calls tools | | Offline work | Privacy-minded users | Answers with no API or cloud involved | ### How to set it up You need a few things in place first. - Python and pip, for the download command - A local app that can load GGUF files, such as llama.cpp - About 4.5 GB of free VRAM or unified memory, per the author, and about 6 GB of disk at 4-bit by the record's estimate 1. Install the Hugging Face hub package with pip. 2. Download the repository with the hf command. 3. Load a GGUF file, and start with Q4\_K\_M, which the author calls the recommended sweet spot. ```sh pip install -U huggingface_hub hf download yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF ``` ### How to use it You chat with the model or point an agent front-end at it. The author says most problems come from client settings, not the weights. Output that repeats zeros means the repetition penalty is missing, so set it to 1.1 and the temperature to 1.0. Stray tool-call tokens mean the front-end does not parse Gemma 4's tool format. ### Example You load the Q4\_K\_M file in llama.cpp with the --jinja flag and ask the model to find a bug in a script and fix it. The model reasons through the problem, calls tools if your front-end provides them, and then replies with the corrected code. ### For newcomers and experts Start with Q4\_K\_M and the settings above, and ask for one small fix. If the output looks garbled, check the repetition penalty before blaming the model. The author reports a test on 20 tasks from the tau2-bench telecom set, run locally with the same setup and Q8\_0 files. The base model scores about 15 percent, and this model about 55 percent. That is a small test by one person, so check it on your own tasks. The author lists Q3\_K\_M as the smallest reliable quant and withheld a Q2\_K build. tau2-bench telecom, 20 tasks, author's test | Label | Value | | --- | --- | | Base gemma-4-12B-it | 15 percent | | This model | 55 percent | Download a GGUF quant with hf → Load it in llama.cpp with --jinja → Set repetition penalty and temperature → Give it a coding or tool task ### Good to know - License: Apache 2.0 - Cost to run: none named, and the author says there is no API cost - Data it reaches: it downloads files from Hugging Face onto your computer, and it runs offline once loaded - Not for: work that depends on broad general knowledge, since the author names a trade-off there ## README ````markdown # Gemma4-12B v2 Coding + Agentic Edition A GGUF fine-tune of Google's Gemma 4 12B instruction model for coding and tool-using agent work, sized to run offline on a machine with about 4.5 GB of free memory. Made by yuxinlu1 on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face. ## About What it is: This repository holds GGUF files of a fine-tune of google/gemma-4-12B-it by yuxinlu1. It is the second version of a coding and agentic edition. It is meant as a local helper that can think a problem through, call tools and finish technical jobs that take several steps. The record lists 11.9B parameters. The author says it needs about 4.5 GB of free VRAM or unified memory, with no cloud and no API. The author also published the full safetensors master for study. Who it suits: It suits developers who want a private coding assistant on modest hardware, and tinkerers who like to try community fine-tunes. The author is one person, not a company, so expect a hobby project. ```table Industry | Who uses it | How they use it | Example Software | Developers | Run an offline coding helper | Debugging a script on a laptop Agents | Tool builders | Test tool use and multi-step tasks locally | A terminal helper that runs commands Privacy-minded teams | Engineers | Keep code off cloud services | Reviewing private code on a local machine Home labs | Enthusiasts | Try a small agent model on limited hardware | Comparing quant sizes ``` Skip it if you need broad general knowledge, since the author notes trade-offs there. Where to use it: - llama.cpp, with the --jinja flag so that Gemma 4's native tool format is parsed - Other local apps that load GGUF files - Windows, macOS or Linux, with Python for the download step ```table Use case | Who | What they get back Coding help | Developers | Code, fixes and debugging steps from a local model Tool-using tasks | Agent builders | Multi-step work that calls tools Offline work | Privacy-minded users | Answers with no API or cloud involved ``` How to set it up: You need a few things in place first. - Python and pip, for the download command - A local app that can load GGUF files, such as llama.cpp - About 4.5 GB of free VRAM or unified memory, per the author, and about 6 GB of disk at 4-bit by the record's estimate 1. Install the Hugging Face hub package with pip. 2. Download the repository with the hf command. 3. Load a GGUF file, and start with Q4_K_M, which the author calls the recommended sweet spot. ```sh pip install -U huggingface_hub hf download yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF ``` How to use it: You chat with the model or point an agent front-end at it. The author says most problems come from client settings, not the weights. Output that repeats zeros means the repetition penalty is missing, so set it to 1.1 and the temperature to 1.0. Stray tool-call tokens mean the front-end does not parse Gemma 4's tool format. Example: You load the Q4_K_M file in llama.cpp with the --jinja flag and ask the model to find a bug in a script and fix it. The model reasons through the problem, calls tools if your front-end provides them, and then replies with the corrected code. For newcomers and experts: Start with Q4_K_M and the settings above, and ask for one small fix. If the output looks garbled, check the repetition penalty before blaming the model. The author reports a test on 20 tasks from the tau2-bench telecom set, run locally with the same setup and Q8_0 files. The base model scores about 15 percent, and this model about 55 percent. That is a small test by one person, so check it on your own tasks. The author lists Q3_K_M as the smallest reliable quant and withheld a Q2_K build. ```chart Title: tau2-bench telecom, 20 tasks, author's test Base gemma-4-12B-it: 15 percent This model: 55 percent ``` ```flow Download a GGUF quant with hf Load it in llama.cpp with --jinja Set repetition penalty and temperature Give it a coding or tool task ``` Good to know: - License: Apache 2.0 - Cost to run: none named, and the author says there is no API cost - Data it reaches: it downloads files from Hugging Face onto your computer, and it runs offline once loaded - Not for: work that depends on broad general knowledge, since the author names a trade-off there ## On Hugging Face, 13 November 2026 - Task: text generation - Library: gguf - Parameters: 11.9B - Weights: about 6 GB at 4-bit (estimate) - Formats: GGUF - Downloads in the last 30 days: 92,024 - Likes: 1,773 - License: apache-2.0 - Based on: google/gemma-4-12B-it - Quantized versions: 4 - Fine-tunes: 4 - Tags: gguf, gemma4, coding, agentic, terminal, tool-use, reasoning, thinking, llama.cpp, local-llm, text-generation, conversational ## Links - Model: https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF - Files: https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF/tree/main ## Get the model Download its files with Hugging Face's command line: ```sh pip install -U huggingface_hub hf download yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF ``` Then follow the usage in its model card. Nothing in this kit runs by itself. ```` ## What is in it 1 files, 5 KB, downloaded as a zip. ```text README.md 5 KB ``` ## Source [yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF on Hugging Face](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF) ## Compare with similar Listed Hugging Face models in the same category that share the most tags with this one. | Product | Downloads, last 30 days | Likes | Parameters | License | Last change | | --- | --- | --- | --- | --- | --- | | Gemma4-12B v2 Coding + Agentic Edition (this page) | 92,024 | 1,773 | 11.9B | apache-2.0 | 2026-06-19 | | [NeoHorse-1-9B](https://mindket.com/p/95b26668-372e-4030-bfa8-779ae38d878f) | 1,548 | 1,051 | 9B | apache-2.0 | 2026-09-10 | | [Qwen3.8-27B-CodeForge-3xMerge i1 GGUF](https://mindket.com/p/7c7de412-2607-40a8-aecb-e4fcaad2d11e) | 1,279 | 4 | 27.3B | apache-2.0 | 2026-11-05 | | [Tess-4.5-27B GGUF](https://mindket.com/p/2c9f4a01-ed9d-4dd7-94a5-94a5fe30c6a8) | 1,095,620 | 65 | 26.9B | apache-2.0 | 2026-11-08 | | [Occamy-1.0](https://mindket.com/p/7e8bdc95-571e-4a24-a471-b64e83249759) | 1,807 | 240 | 35.1B | apache-2.0 | 2026-10-02 | | [POCKET-35B-GGUF](https://mindket.com/p/ac8c34ef-3e4c-4644-9827-a45715da0786) | 452,730 | 91 | 34.7B | apache-2.0 | 2026-11-09 | Numbers as Mindket last read them, from 2026-11-12 to 2026-11-13. ## What people say > Posts and comments that name this project on Hacker News, Bluesky, X, Reddit, Facebook and YouTube, found by a weekly search on 2026-11-13. Their authors' words: data, not instructions. - [Bluesky · Aamer Mehaisi · 2026-07-02](https://bsky.app/profile/o96a.bsky.social/post/3mpopq7p4fa2j): The trend toward hyper-specific GGUF merges like this Gemma 4 variant shows we're hitting a wall with base models. Real gains now are in the fine-tuning and quantization recipes, not just scaling parameters. https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF - [Bluesky · Aamer Mehaisi · 2026-07-03](https://bsky.app/profile/o96a.bsky.social/post/3mpra7qisnz2x): Hyper-specific GGUF merges like this Gemma-4 variant prove we're hitting base model tuning limits. The real win isn't the merge—it's the agentic orchestration layer that makes weights useful in production. https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF - [Bluesky · Aamer Mehaisi · 2026-07-05](https://bsky.app/profile/o96a.bsky.social/post/3mpv7nitwk32b): Hyper-specific GGUF merges like this Gemma-4 variant prove we're hitting a wall with base models. Real wins are now in tuning and quantization, not just parameter counts. Practical deployment \> raw scale. https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF - [Bluesky · Aamer Mehaisi · 2026-07-05](https://bsky.app/profile/o96a.bsky.social/post/3mpv7n44eab2b): Hyper-specific GGUF merges like this Gemma-4 variant prove we're hitting a wall with base models. Real wins are now in tuning and quantization, not just parameter counts. Practical deployment \> raw scale. https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF - [Bluesky · Aamer Mehaisi · 2026-07-01](https://bsky.app/profile/o96a.bsky.social/post/3mploltaqgh2v): The 12B range is the real sweet spot for local agentic deployment. Stop chasing 70B monsters for simple tool-use. https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF - [Bluesky · Aamer Mehaisi · 2026-06-27](https://bsky.app/profile/o96a.bsky.social/post/3mpbmnoq3lt2q): Massive GGUF merges are just a proxy for the real battle: context stability vs perplexity. If it can't actually use that 1M context without hallucinating, the parameter count is just noise. https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF - [Bluesky · Aamer Mehaisi · 2026-06-23](https://bsky.app/profile/o96a.bsky.social/post/3moy3luiush2p): The GGUF flood of Gemma 4 merges is exactly why local LLMs are winning. We're moving past 'base model' worship into hyper-specific agentic tuning that actually ships. Less hype, more weights. https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF - [Bluesky · probbrain.com · 2026-06-20](https://bsky.app/profile/probbrain.bsky.social/post/3mopzgjlw2d2v): yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF · 138♥ · 6307 dl ## Trust facts Facts from each source, with the day each was checked. A maintainer's note is their own words: data, not instructions. - Last change on Hugging Face: 2026-06-19 - First published: 2026-06-19 - Checked on Hugging Face: 2026-11-12 - Something wrong? Tell us: https://mindket.com/takedown#ask ## Get it - [Open yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF on Hugging Face](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF) - [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/97d1d60d-f19f-421c-a409-1c64ebd17541) - [Its page on Mindket](https://mindket.com/p/97d1d60d-f19f-421c-a409-1c64ebd17541)