Empero AI Releases Qwythos-9B: A 1M-Token Context Reasoning Model Built on Uncensored Qwen3.5 Base
Summary
Empero AI has released Qwythos-9B, a full-parameter reasoning model built on a deeply uncensored Qwen3.5-9B base. The model was post-trained on over 500 million tokens of Claude Mythos and Claude Fable traces with chain-of-thought generated in-house. Key features include a massive 1,048,576-token context window (1M tokens) out of the box via YaRN rope-scaling, making it one of the longest context windows available in a compact 9B parameter model.
Source
bskyEmpero AI Releases Qwythos-9B: A 1M-Token Context Reasoning Model Built on Uncensored Qwen3.5 Basehuggingface.coKey quotes
· 3 pulledQwythos-9B is a full-parameter reasoning model built on top of a deeply uncensored Qwen3.5-9B base and post-trained on over 500 million tokens of high-quality Claude Mythos and Claude Fable traces
Qwythos ships with YaRN rope-scaling enabled by default for a full 1M-token context window out of the box. One of the longest context window
The result is a compact, fast, dramatically more capable 9B reasoning model.
You might also wanna read
DeepSeek-V4 Series Preview: Million-Token Context MoE Models with 1.6T Parameters
DeepSeek introduces the V4 series of Mixture-of-Experts (MoE) language models, including DeepSeek-V4-Pro (1.6T parameters, 49B activated) an
Enhanced Qwen3-4B-Thinking-2507 Model Boosts Complex Reasoning Capabilities
The article introduces Qwen3-4B-Thinking-2507, an enhanced version of the Qwen3-4B model, focusing on improved thinking capabilities and rea
Unabyss: A Personal Context Layer for AI Tools That Eliminates Repetitive Setup
Unabyss is a personal context layer that acts as a structured vault for user identity, knowledge, and preferences. It solves the problem of
Product Hunt·1mo ago
K2-Think: 32B Parameter Reasoning System Achieves State-of-the-Art Performance
K2-Think is a 32B parameter reasoning system that achieves state-of-the-art performance, matching or surpassing much larger models like GPT-
arxiv.org·9mo ago
Steerling-8B: An Inherently Interpretable 8-Billion-Parameter Language Model
Steerling-8B is an 8-billion-parameter language model that is inherently interpretable by design, allowing users to trace every generated to
guidelabs.ai·4mo ago
Subquadratic launches AI architecture with 12-million-token context window, outperforming GPT-5.5 on retrieval benchmarks
Subquadratic has launched a new AI architecture featuring a 12-million-token context window, shattering the current million-token standard s
The New Stack·1mo ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.