interpolators/FableOpus-9B-TIES

VISIONConcurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 21, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

interpolators/FableOpus-9B-TIES is a 9 billion parameter Qwen3.5-based model merged using a delta_tiesish method, combining Fable 5 distillation with Claude Opus reasoning distilled checkpoints. This bf16 safetensors merge is designed for general chat, code assistance, and tool-use style prompting. It excels in reasoning-heavy tasks by integrating agentic capabilities and stabilized reasoning from its source signals.

Loading preview...

Model Overview

interpolators/FableOpus-9B-TIES is a 9 billion parameter language model built upon the Qwen3.5-9B architecture. It is a bf16 safetensors merge that leverages a delta_tiesish method to combine distinct capabilities from its source models. The merge integrates the agentic and tool-use characteristics derived from Fable 5 distillation with the enhanced reasoning stabilization found in Claude Opus distilled checkpoints.

Key Capabilities

  • Reasoning Enhancement: Incorporates reasoning capabilities from high-download Claude Opus distilled checkpoints.
  • Agentic & Tool-Use: Designed to support agentic workflows and tool-use style prompting, stemming from Fable 5 distillation traces.
  • Code Assistance: Suitable for generating and assisting with code-related tasks.
  • General Chat: Capable of handling general conversational interactions.

Merge Recipe

The model was created by merging three distinct checkpoints with specific weights:

  • empero-ai/Qwable-9B-Claude-Fable-5: 0.52
  • Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled: 0.33
  • Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2: 0.15

Intended Use Cases

This model is primarily intended for:

  • General chat applications.
  • Assisting with code generation and understanding.
  • Experiments involving tool-use style prompting.
  • Tasks requiring strong reasoning capabilities.