masterzm9/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-safetensors
The masterzm9/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-safetensors model is a 9 billion parameter multimodal large language model from the Qwen3.5 family, developed by Qwen. It features a unified vision-language foundation, an efficient hybrid architecture with Gated Delta Networks and Mixture-of-Experts, and expanded global linguistic coverage to 201 languages. This model excels in multimodal reasoning, coding, agentic tasks, and visual understanding, supporting a native context length of 262,144 tokens and extensible up to 1,010,000 tokens.
Loading preview...
Qwen3.5-9B: A Multimodal Agentic LLM
This model is a 9 billion parameter variant of the Qwen3.5 series, developed by Qwen, and is a safetensors conversion of the HauhauCS aggressive fine-tune. It represents a significant advancement in foundation models, integrating multimodal learning, architectural efficiency, and scalable reinforcement learning.
Key Capabilities
- Unified Vision-Language Foundation: Achieves strong performance across reasoning, coding, agentic tasks, and visual understanding benchmarks by early fusion training on multimodal tokens.
- Efficient Hybrid Architecture: Utilizes Gated Delta Networks and sparse Mixture-of-Experts for high-throughput inference with minimal latency.
- Scalable RL Generalization: Trained with reinforcement learning across millions of agent environments for robust real-world adaptability.
- Global Linguistic Coverage: Supports 201 languages and dialects, enabling broad deployment with cultural and regional understanding.
- Extended Context Length: Natively handles up to 262,144 tokens, extensible to 1,010,000 tokens using YaRN scaling techniques.
- Tool Calling: Demonstrates strong capabilities in tool use, with recommended integration via Qwen-Agent and Qwen Code.
Performance Highlights
The Qwen3.5-9B model shows competitive performance across various benchmarks, including:
- Language: Achieves 82.5 on MMLU-Pro, 88.2 on C-Eval, and 91.5 on IFEval.
- Vision Language: Scores 78.4 on MMMU, 70.1 on MMMU-Pro, and 85.7 on Mathvista (mini).
- Agentic Tasks: Excels in general agent benchmarks like BFCL-V4 (66.1) and TAU2-Bench (79.1).
- Multilingualism: Performs well on MMMLU (81.2) and MMLU-ProX (76.3).
Should I use this for my use case?
This model is ideal for applications requiring advanced multimodal understanding, complex reasoning, code generation, and agentic capabilities, especially where long context windows and broad language support are crucial. Its optimized architecture makes it suitable for scenarios demanding efficient inference. Developers building applications that leverage vision, video, and text inputs, or those requiring robust tool-use integration, will find this model particularly effective. Consider its strong performance in instruction following and general agent tasks for building intelligent assistants or automated workflows.