Sodnom-AI/huihui-qwen36-abliterated-chat

Huihui Qwen3.6-35B-A3B Claude-4.7-Opus abliterated — chat

A static, in-browser chat for huihui-ai/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated, the abliterated (refusal-removed) build of a Qwen3.6-35B-A3B model distilled on Claude-Opus reasoning traces.

How it works

There is no server and no GPU here — the page runs entirely in your browser and calls the model through Hugging Face Inference Providers (provider featherless-ai), which serves it at near-full precision. That means zero hosting cost and you only pay per token for what you actually generate, billed to your own Hugging Face account.

To use it: click ⚙️ Token & settings, paste a Hugging Face token that has the “Make calls to Inference Providers” permission (create one here — type Fine-grained), and Save. The token is stored in your browser only (localStorage) and sent directly to Hugging Face; it is never uploaded to this Space. Free accounts get a small monthly inference credit; beyond that you need HF PRO or a payment method.

Prefer to run it fully locally instead? Use the GGUF quants (Q2_K … f16, +vision mmproj): huihui-ai/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-MTP-GGUF (needs llama.cpp ≥ b7990, which added the qwen35moe architecture).

⚠️ Usage warnings (from the model card)

This model's safety filtering has been significantly reduced. It may generate sensitive, controversial, or inappropriate content. Not suitable for all audiences; intended for research, testing, and controlled environments. You are solely responsible for how you use the outputs and for complying with local laws. No default safety guarantees are provided by the model author.