nightmedia/Qwen3.5-9B-Brainwaves

VISIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 29, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

nightmedia/Qwen3.5-9B-Brainwaves is a 9 billion parameter experimental merge model based on the Qwen3.5 architecture, combining schneewolflabs/Wichtelchen-Qwen3.5-9B and nightmedia/Qwen3.5-9B-Holodeck-Lounge. This model is designed for complex tasks requiring both precise technical understanding and creative, philosophical narrative capabilities, excelling in agentic programming and rich storytelling. It features a 32768 token context length and demonstrates improved perplexity and benchmark scores over its base components.

Loading preview...

nightmedia/Qwen3.5-9B-Brainwaves: A Synergistic Merge Model

This 9 billion parameter model is an experimental merge of two distinct Qwen3.5-9B variants: schneewolflabs/Wichtelchen-Qwen3.5-9B and nightmedia/Qwen3.5-9B-Holodeck-Lounge. The merge was intentionally weighted (1.4 to 0.6) to combine their strengths, resulting in a model that excels in both technical precision and creative depth.

Key Capabilities

  • Advanced Technical Reasoning: Inherits robust capabilities for handling engineering tasks, code generation, tool utilization, and syntax logic, fine-tuned on custom DPO datasets for programming and delegation.
  • Creative & Philosophical Depth: Integrates a rich blend of creative writing, deep philosophy, and agent capabilities, drawing from models like Claude-4.6-OS, The-Bradbury, and Mark-Twain.
  • Improved Coherence: Demonstrates a lower perplexity (4.292) compared to its parent models, indicating cleaner and more coherent output.
  • Enhanced Benchmarks: Shows improved performance across various benchmarks, including an ARC score of 0.687 and a BoolQ score of 0.904.
  • Multi-Persona Interaction: Capable of holding advanced technical conversations while filtering understanding through rich, episodic narratives, allowing for multi-turn, multi-modal interactions.

Ideal Use Cases

  • Agentic Programming: Excellent for tasks requiring code generation, system architecture understanding, and tool orchestration.
  • Creative Content Generation: Suited for generating rich, philosophical, and imaginative narratives.
  • Complex Problem Solving: Effective in scenarios demanding both precise technical solutions and nuanced, contextual understanding.
  • Interactive Storytelling: Can comfortably engage in multi-persona, multi-modal narrative engines, blending technical and creative elements.

Finetunes
3 models