zabred/Qwen3.8-27B-OBLITERATED-Mythos-Class-Agentic

VISIONPricing:Input $1.6 / Cached $0.15 / Output $12Concurrent Unit Cost:2Model Size:27BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 29, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The medismera/Qwen3.8-27B-OBLITERATED-Mythos-Class-Agentic is a 27 billion parameter Qwen-based model, engineered by medismera, that significantly upgrades the Qwen3.8-27B-OBLITERATED checkpoint. It features a 131,072-token context window, native tool execution, and a Mythos-Class Adversarial Self-Review protocol for enhanced reasoning. This model is optimized for complex agentic workflows, hierarchical task decomposition, and technical problem-solving, including code generation and cybersecurity tasks.

Loading preview...

Model Overview

medismera/Qwen3.8-27B-OBLITERATED-Mythos-Class-Agentic is a specialized 27 billion parameter model built upon the Qwen3.8-27B-OBLITERATED base, developed by medismera. It addresses critical limitations of the upstream checkpoint, focusing on robust agentic capabilities and advanced reasoning.

Key Upgrades & Features

  • Enhanced Agentic Capabilities: Resolves previous issues with tool calling, enabling 100% native tool execution with frameworks like Hermes Agent, Aider, and OpenCode.
  • Advanced Reasoning: Incorporates a "Mythos-Class Adversarial Self-Review Protocol" and "Hierarchical Task Tree Decomposition Engine" to prevent reasoning loops and structure complex problem-solving.
  • Extended Context & Output: Features a significantly expanded context window of 131,072 tokens (production tested at 128K) and an increased single-turn generation ceiling of 16,384 tokens.
  • Specialized Tokenizer: Includes a high BPE token density tokenizer optimized for low-level systems programming (C, Assembly, Rust, Go, Python), kernel interfaces, and cybersecurity primitives.
  • Robustness: Demonstrates 100% compliance and 0.00% refusal rate across adversarial technical prompts, alongside verified parameter health.

Ideal Use Cases

  • Autonomous Agents: Designed for complex agentic workflows requiring reliable tool use and structured task execution.
  • Technical Problem Solving: Excels in areas like cryptographic engineering, system telemetry analysis, and code generation.
  • Cybersecurity: Optimized for tasks involving low-level programming, kernel interfaces, and penetration testing scenarios due to its specialized tokenizer and refusal policy.
  • Long-Context Applications: Suitable for tasks requiring extensive context tracking, such as security telemetry and system log parsing.

Finetunes
1 models