januschoy/druckenmiller-1.5b-v2

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:1.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jul 20, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

januschoy/druckenmiller-1.5b-v2 is a 1.5 billion parameter experimental language model based on Qwen2.5-1.5B-Instruct, fine-tuned for macro trading cognitive sparring in the style of Stanley Druckenmiller. It features a 32768 token context length and is specifically designed for research and educational demonstrations related to financial market analysis. This model is optimized for generating insights on macro trading, offering a specialized application within the financial domain.

Loading preview...

Druckenmiller 1.5B v2: Macro Trading Cognitive Sparring Model

This experimental model, developed by januschoy, is a specialized 1.5 billion parameter language model built upon the Qwen/Qwen2.5-1.5B-Instruct base. It has been fine-tuned using QLoRA (r=16, 4 epochs) on approximately 284 ShareGPT conversations, with about 28% being multi-turn interactions. The primary goal of Druckenmiller 1.5B v2 is to serve as a cognitive sparring partner for macro trading, emulating the style of Stanley Druckenmiller.

Key Capabilities

  • Specialized Financial Domain: Focused on generating insights and responses related to macro trading and financial market analysis.
  • Qwen2.5 Base: Leverages the capabilities of the Qwen2.5-1.5B-Instruct architecture.
  • Experimental Preview: Currently in an experimental preview state, intended for research and educational demonstrations.

Known Limitations

  • Multi-turn Coherence: May occasionally deviate from the topic during multi-turn follow-up questions.
  • Capacity Constraints: Due to its 1.5B parameter size, it might hallucinate macro-level details.
  • Performance: Public demos running on free CPU spaces may experience high latency and cold start issues.

Good For

  • Research and Education: Ideal for academic study and demonstrating AI applications in financial contexts.
  • Macro Trading Simulation: Useful for exploring hypothetical scenarios and gaining perspectives on macro trading strategies.

Disclaimer: This model is for research and educational demonstration purposes only and does not constitute investment advice. Financial markets involve risks, and users are responsible for their own decisions. The model weights follow an Apache-2.0 license, with adherence to Qwen2.5 terms.

Finetunes
1 models