Anthropic · Alibaba · Claude · Llama · Decrypt
It runs in GGUF format—the compressed, consumer-friendly file type that works with LM Studio or llama.cpp—and fits in roughly 16.5
Compiled by KHAO Editorial — aggregated from 4 sources. See llms.txt for citation guidance.
✓ KHAO Verified
It sends nothing to Anthropic's servers, which matters given that Fable 5 required mandatory 30-day data retention on all traffic, even for enterprise customers who previously had zero-retention agreements.
Key facts
- It's a full fine-tune of Alibaba's Qwen3.6-27B base, built by developer Mia (Mia-AiLab on Hugging Face) on a dataset of Fable 5-style reasoning examples
- The reporter has trained Qwen 3.6 27b with Fable 5 reasoning
- It runs in GGUF format—the compressed, consumer-friendly file type that works with LM Studio or llama.cpp—and fits in roughly 16.5 GB in its Q4 quantized build
- A similar approach drove Qwopus —the Claude Opus 4.6 local distillation—though that project focused on chain-of-thought reasoning traces
Summary
Qwable 27B is a full fine-tune of Alibaba's Qwen3.6-27B, trained on a Fable 5-style reasoning dataset, designed to replicate the structured, deliberate thinking style of Anthropic's newest flagship model. The abliterated version removes the model's built-in refusal behavior by surgically modifying its weights using llama.cpp's cvector-generator. Both models run locally, cost nothing per query, and require neither Anthropic's API nor its mandatory policies. Anthropic spent last week apologizing for Fable 5's invisible safeguards, and then the U.S. government ordered the model pulled for all foreign nationals over a disputed jailbreak finding. A few days later, a developer on Hugging Face uploaded a model that used Fable’s reasoning to guide a local model—and now even your potato PC can run a better model.