Back to News

### TL;DR

Qwythos-9B-Claude-Mythos-5-1M-GGUF is a 9-billion parameter reasoning model developed by Empero AI. It is post-trained on over 500 million tokens of high-quality Claude Mythos and Claude Fable traces, utilizing Empero AI's internal 'rethink' tool for chain-of-thought generation. This model outperforms the base Qwen3.5-9B model, achieving a +34 point improvement on the MMLU benchmark, +30 points on gsm8k-strict, and +19 points on gsm8k-flex. It supports native function calling as per the Qwen3.5 specification and offers a 1,048,576-token (1M) context window through YaRN rope-scaling. The model is available under the Apache-2.0 license and is compatible with various GGUF runtimes, including llama.cpp, Ollama, LM Studio, jan, and KoboldCpp.

Key Insights & Metrics

Key Features

• 9-billion parameter reasoning model
• Post-trained on over 500 million tokens of high-quality Claude Mythos and Claude Fable traces
• Outperforms base Qwen3.5-9B model on multiple benchmarks
• Supports native function calling as per Qwen3.5 specification
• Offers 1,048,576-token (1M) context window through YaRN rope-scaling
• Available under Apache-2.0 license
• Compatible with various GGUF runtimes, including llama.cpp, Ollama, LM Studio, jan, and KoboldCpp

Discussion