AIMarkTechPostabout 19 hours ago
657MB Thinking Model Fine-Tuned on Claude Traces
3 min read
A developer fine-tuned OpenBMB's MiniCPM5-1B on Claude Fable 5 traces to create a 657MB local thinking model. The GGUF model runs fully offline in llama.cpp, Ollama, LM Studio, jan, and KoboldCpp with no API calls. This is supervised fine-tuning on generated outputs, not weight-level distillation.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyAnthropic
