---

base_model:

• Kwaipilot/KAT-Coder-V2.5-Dev
• Accio-Lab/occamy-1.0
• EryriLabs/Ornith-1.5-35B-A3B-BigBang-MTP-GGUF

base_model_relation: merge

library_name: transformers

pipeline_tag: text-generation

tags:

• qwen3.5
• agent
• coding
• code
• kat-coder
• ornith
• occamy
• reasoning
• tool-use
• safetensors
• gguf
• apex
• imatrix
• llama.cpp

---

Qwen3.6-Qworus-35B-A3B-APEX-GGUF

What is this?

This is a 50/30/20 DARE-TIES merge of Kwaipilot/KAT-Coder-V2.5-Dev, Accio-Lab/occamy-1.0, and EryriLabs/Ornith-1.5-35B-A3B-BigBang-MTP-GGUF, respectively.

The Ornith/BigBang merge is used as the base model, so this model should inherit the same configuration and chat template.

The result is a hybrid model for coding, general agent work, planning, designing, and conversation.

---

Quantization

Weights are provided quantized with APEX, which preserves sensitive MoE tensors. I-Quality is the best size for high quality coding and reasoning, but I may add other sizes later.

The importance matrix was calibrated on about 50% random English text snippets from multiple domains, and 50% tokens generated during real long-context agent sessions with KAT-Coder and Ornith-1.5.

---

Merge Recipe

A mergekit-style YAML was used to produce this model. However, the actual merge was performed by Alloy, rather than Mergekit, as the latter kept producing a broken model which could not be quantized.

merge_method: dare_ties
base_model: EryriLabs/Ornith-1.5-35B-A3B-BigBang-MTP-GGUF
tokenizer_source: EryriLabs/Ornith-1.5-35B-A3B-BigBang-MTP-GGUF
dtype: bfloat16
out_dtype: bfloat16

models:
  - model: Kwaipilot/KAT-Coder-V2.5-Dev
    parameters:
      weight: 0.50
  - model: Accio-Lab/occamy-1.0
    parameters:
      weight: 0.30
  - model: EryriLabs/Ornith-1.5-35B-A3B-BigBang-MTP-GGUF
    parameters:
      weight: 0.20
parameters:
  int8_mask: true
  normalize: true
  lambda: 1.05
  density:
    - value: 0.7
    - filter: ".*(mlp\\.gate|mlp\\.shared_expert|self_attn\\.[qkvo]_proj).*"
      value: 1.0
  epsilon:
    - value: 0.12
    - filter: ".*(mlp\\.gate|mlp\\.shared_expert|self_attn\\.[qkvo]_proj).*"
      value: 0.0

Most of the model's weights were merged with DARE-TIES with a density of 0.7 and epsilon of 0.12, except for the router gates, shared experts, and self attention layers which are protected from sparsification by using a density of 1.0 and epsilon of 0.0, instead.