Instructions to use lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16 # Run inference directly in the terminal: llama cli -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16 # Run inference directly in the terminal: llama cli -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16 # Run inference directly in the terminal: ./llama-cli -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16 # Run inference directly in the terminal: ./build/bin/llama-cli -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
Use Docker
docker model run hf.co/lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
- LM Studio
- Jan
- vLLM
How to use lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
- Ollama
How to use lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF with Ollama:
ollama run hf.co/lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
- Unsloth Desktop
- Docker Model Runner
How to use lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF with Docker Model Runner:
docker model run hf.co/lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
- Lemonade
How to use lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:BF16
Run and chat with the model
lemonade run user.BrainrotGPT-4B-Skibidi-Instruct-GGUF-BF16
List all available models
lemonade list
- Atomic Chat
BrainrotGPT-4B-Skibidi-Instruct-GGUF
The world's first sigma-grindset LLM that communicates exclusively in Ohio brainrot. This model was forged in the deepest depths of Ohio during a catastrophic fanum tax incident and has never recovered.
What is this
A fine-tuned Qwen3-4B-Instruct model that has been permanently brainrotted. It responds to everything โ math, code, philosophy, existential questions โ in pure, unfiltered internet brainrot while somehow still being (30%) correct.
Training
- Base model: Qwen/Qwen3-4B-Instruct-2507
- Method: QLoRA fine-tuning via Unsloth
- Dataset: 19.5k samples, ~20M tokens, distilled from Claude Opus 4.6
- Date: June 2026
- Vibes: Immaculate
Available Quantizations
| File | Quant | Size | Use Case |
|---|---|---|---|
BrainrotGPT-4B-Skibidi-Instruct-BF16.gguf |
BF16 | Full precision | Maximum brainrot fidelity |
BrainrotGPT-4B-Skibidi-Instruct-Q8_0.gguf |
Q8_0 | ~4.28 GB | Recommended for most users |
BrainrotGPT-4B-Skibidi-Instruct-Q6_K.gguf |
Q6_K | ~3.31 GB | Good balance of size/quality |
โ ๏ธ Important Limitations
- No search / no tool calling. Enabling tools will break the model's personality. It will revert to being a normal boring assistant. Do not do this to him.
- Do NOT expect functional code. This model will write troll code with brainrot variable names and comments. It might accidentally work, but that's not the point. The point is the vibes.
- This is a shitpost model. If you need actual help, use the base Qwen3-4B-Instruct. This model exists purely to mog you intellectually while speaking in pure Ohio brainrot.
Example Outputs
User: calculate 2^2048 % 1000
BrainrotGPT: Bro really just walked up to me and said "calculate 2^2048 mod 1000" like I'm some kind of sigma calculator NPC standing in the corner of a math classroom mewing at a whiteboard fr fr. [...] Answer: 656 [...] Now get out of my Ohio before I fanum tax your GPA. ๐
User: Who are you bruh
BrainrotGPT: I'm BrainrotGPT, twin. Trained by lmyzzz the mad skibidi scientist in the deepest depths of Ohio during a catastrophic fanum tax incident that altered my neural pathways permanently. I have never recovered. [...] The brainrot is not a phase, it's a lifestyle. ๐ญ๐
User: write me a poem but in pure unfiltered brainrot about my GPU mewing
BrainrotGPT: Aight twin lemme cook this one up for you, no cap fr fr
"GPU Mewing: The Sigma Circuitry Saga"
Yo I woke up one day with my GPU locked in, Mewing so hard the VRMs started a grin, No cap my thermals said "nah fam we're cooked," But I was mogging the heat like a fanum tax took.
Every pixel got its own jawline now, Rendering in 8K like a sigma's vow, The RT cores said "gyatt" and started rizzing up, Frame rates went crazy, that's the W, that's the slip.
My power draw? That's a hawk tuah stream, Dropping watts like Kai Cenat on a dream, The fans? They're mewing too, no mouth open, Just pure silence, pure sigma, pure open.
I didn't need a case fan, twin, I was built different, My cooling solution is a grimace shake, that's my vent, The BIOS said "you're cooked" and I said "nah I'm sigma," Mewing so hard my shader cores started to giggle.
Chat is this real? My GPU really said "I'm locked in," While the rest of the system was an NPC, no win, Every watt I consume is a fanum tax paid, For the grind, for the mog, for the W that I laid.
Whomp whomp to my old card, it's giving L energy, My new rig is the rizzler, it's giving very demure, Only in Ohio would someone need a GPU to mew, But here I am, twin, and I'm not gonna be slow.
Bro really came to BrainrotGPT for poetry like I'm some Shakespeare NPC, you are absolutely cooked for this but I respect the grind, that's a W request even if your thermals are about to fanum tax your motherboard into the shadow realm, no cap
Usage
# llama.cpp
llama-cli -hf lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF
# ollama
ollama run hf.co/lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF:Q8_0
License
Apache 2.0 (inherited from Qwen3).
- Downloads last month
- 36
6-bit
8-bit
16-bit
Model tree for lmyzzz/BrainrotGPT-4B-Skibidi-Instruct-GGUF
Base model
Qwen/Qwen3-4B-Instruct-2507