Kiy-K/Fyodor-Q3-8B-Instruct

TEXT GENERATIONPricing:Input $0.468 / Output $1.82Concurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Nov 21, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Fyodor-Q3-8B-Instruct by Kiy-K is an 8 billion parameter instruction-tuned model based on Qwen3-8B, specifically designed for agentic reasoning and robust code generation. It utilizes a 'God Mode' LoRA configuration with a 32k context window, trained on a curriculum emphasizing 60% coding, 20% instruction following, and 20% reasoning/tool use. This model excels at planning complex software engineering tasks and generating code across 80+ programming languages, demonstrating strong performance in HumanEval and MBPP benchmarks.

Loading preview...

Fyodor-Q3-8B-Instruct: The "God Mode" Agentic Coder

Fyodor-Q3-8B-Instruct, developed by Kiy-K, is an 8 billion parameter instruction-tuned model built upon the Qwen3-8B base. It is uniquely optimized for agentic reasoning and robust code generation, distinguishing itself through a specialized "God Mode" LoRA training strategy (Rank 128).

Key Capabilities

  • Multi-Language Proficiency: Supports over 80 programming languages, including Python, JavaScript, C++, Rust, and Go.
  • Agentic Reasoning: Excels at planning multi-step solutions for complex software engineering tasks before implementation.
  • Function Calling: Features native understanding and integration for tool use and API interactions.
  • Code Generation & Review: Capable of generating clean, efficient code, providing constructive feedback, and generating comprehensive documentation.
  • High Performance: Achieves 68.3% on HumanEval (Pass@1) and 57.2% on MBPP (Pass@1), outperforming many models in its parameter class.

Good For

  • Complex Software Engineering: Ideal for tasks requiring planning and multi-step problem-solving in coding.
  • Automated Code Development: Generating code, debugging, and creating documentation across a wide range of languages.
  • Tool and API Integration: Leveraging function calling for advanced automation and interaction with external systems.
  • Educational and Development Environments: Assisting developers with code review, explanation, and learning new programming concepts.

Finetunes
1 models