Find a model for your next idea

Explore AI models.

Browse standard publisher releases and community-modified models. Find a purpose, compare formats, and check what fits your setup.

Browse models Setup guidance Full model directory →

Two collections. One directory.

Standard means publisher releases without a listed abliteration modification, including publisher-provided quantizations. It is not a guarantee about safety or refusal behavior. Abliterated means community modifications designed to reduce learned refusal behavior; this does not guarantee accuracy or better capability.

About abliteration

Curated snapshot checked November 1, 2026. Standard picks were observed on Hugging Face’s Trending list; rankings change. Sources reviewed, not inference-tested. RAM estimates on the original Qwen3 cards are planning guidance, not measured minimums. Download size is not runtime memory usage.

Hugging Face Trending ↗Huihui publisher ↗

Loading model directory…

Run your first model.

With Ollama

Install Ollama for your operating system using its official guide.
For an Ollama-ready card, start Ollama, open a terminal, and paste its run command above. Other formats need the workflow in their model card. The first run downloads the model.
Try a simple prompt such as “Help me outline a weekend project.” Compare the result with the base model if you want to understand the changes.

Ollama setup guide

Prefer a desktop interface?

LM Studio can run compatible GGUF files. Open a model card, look for linked GGUF conversions, and check the conversion author and model version before downloading. The original Safetensors weights are a different format from GGUF.

Choose a quantization that fits your available memory. If loading fails, try a smaller model or shorter context.

GGUF with LM Studio

Compare notes with FRONTIERS.

Share the exact model tag, your RAM/GPU, context setting, and what worked. That helps other members choose a setup.

Ask or share in Discord