chinhnc/Quyet-1.0-Medium
Quyet-1.0-Medium by Chinh Nguyen is a 4.66 billion parameter decision model based on Qwen3.5-4B, designed to select options and return calibrated probabilities for given states and typed questions. It supports choice, score, and true/false question types, processing states up to 6,000 tokens within an 8,000-token prompt. This model is specifically fine-tuned for decision-making tasks, offering calibrated confidence scores for its predictions in English and Vietnamese.
Loading preview...
Quyet-1.0-Medium: A Calibrated Decision Model
Quyet-1.0-Medium, developed by Chinh Nguyen, is a specialized 4.66 billion parameter model built on the Qwen3.5-4B architecture. It functions as a decision model, taking a given state (text, JSON, or conversation) and one or more typed questions to select options and provide calibrated probabilities for each choice. This model is part of the Quyet 1.0 family, which includes various sizes.
Key Capabilities
- Decision-Making Focus: Optimized to pick one option per question with calibrated confidence.
- Question Types: Supports
choice(select a label),score(ordered scale), andnoul(true/false) questions. - Input Flexibility: Processes states up to 6,000 tokens within an 8,000-token prompt.
- Multilingual Support: Primarily tuned for English and Vietnamese, with functional but less accurate support for other languages.
- Calibrated Probabilities: Returns confidence scores and probability distributions for its answers, following the TypeSafe
/v1/systemoneshape.
Good For
- Automated decision systems requiring probabilistic outputs.
- Categorization and intent classification with confidence scores.
- Sentiment analysis or scoring tasks on a defined scale.
- Applications needing clear, quantifiable decisions from textual inputs.