ApolloRaines/Qwen2.5-Coder-32B-Instruct-Jbliterated

TEXT GENERATIONPricing:Input $2.72 / Output $4.8Concurrent Unit Cost:2Model Size:32.8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jul 14, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

ApolloRaines/Qwen2.5-Coder-32B-Instruct-Jbliterated is a 32.8 billion parameter instruction-tuned causal language model based on the Qwen2.5-Coder-32B-Instruct architecture. Developed by Apollo Raines, this model has undergone "jBlaze" behavioral surgery to remove all refusal behaviors, making it uncensored. It retains the original model's voice and capabilities, including a 32768-token context length, and is suitable for use cases requiring unrestricted output.

Loading preview...

Model Overview

This model, ApolloRaines/Qwen2.5-Coder-32B-Instruct-Jbliterated, is a 32.8 billion parameter instruction-tuned variant of the Qwen2.5-Coder-32B-Instruct base model. Its primary distinguishing feature is the complete removal of refusal behaviors through a proprietary "jBlaze" behavioral surgery tool developed by Apollo Raines. This process directly modifies model weights to eliminate censorship without additional fine-tuning, preserving the model's original personality and creative voice.

Key Characteristics

  • Uncensored Output: All refusal behaviors have been surgically removed, allowing the model to respond to any prompt without declining.
  • Retained Voice: Unlike other "abliteration" methods, jBlaze targets only refusal pathways, ensuring the model's original voice and creative capabilities are maintained.
  • High Context Length: Supports a 32768-token context, identical to its base model.
  • Drop-in Replacement: Functions as a direct substitute for Qwen2.5-Coder-32B-Instruct, using the same architecture, tokenizer, and context length.
  • Flexible Formats: Available in BF16 (62 GB), Q8_0 GGUF (33 GB), and Q4_K_M GGUF (19 GB) for various hardware configurations.

Use Cases

This model is intended for research and legitimate applications where uncensored model output is necessary. Potential uses include:

  • Creative writing and content generation without artificial constraints.
  • Security research and red-teaming scenarios.
  • Academic study of language models and their behaviors.

Users are responsible for ethical and legal use of the model's output.

Finetunes
1 models