Zhongzhi1228/Qwen3.5-27B-RL

VISIONPricing:Input $1.6 / Cached $0.15 / Output $12Concurrent Unit Cost:2Model Size:27BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 3, 2026License:cc-by-4.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Zhongzhi1228/Qwen3.5-27B-RL is a 27.8 billion parameter Qwen3.5 causal language model, developed by Zhongzhi1228, featuring a vision encoder and an extended 262,144-token context length. This model is a reinforcement learning checkpoint derived from Qwen/Qwen3.5-27B, specifically associated with the Recursive Task Synthesis project. It is designed for advanced multimodal language tasks, leveraging its large context window and RL fine-tuning for complex sequence understanding and generation.

Loading preview...

Overview

Zhongzhi1228/Qwen3.5-27B-RL is a 27.8 billion parameter multimodal causal language model, fine-tuned using reinforcement learning. Derived from the base Qwen/Qwen3.5-27B model, it incorporates a vision encoder and boasts an exceptionally long context length of 262,144 tokens. This model is specifically developed as a checkpoint for the Recursive Task Synthesis project, indicating its optimization for tasks requiring complex, multi-step reasoning and synthesis.

Key Capabilities

  • Multimodal Understanding: Integrates a vision encoder, allowing for processing and understanding of both text and visual inputs.
  • Extended Context Window: Features a 262,144-token context length, enabling the model to handle very long sequences and complex, multi-turn conversations or documents.
  • Reinforcement Learning Fine-tuning: Benefits from RL optimization, likely enhancing its ability to follow instructions, generate coherent and relevant responses, and perform well on specific task objectives.
  • Recursive Task Synthesis: Directly associated with the Recursive Task Synthesis project, suggesting specialized performance in tasks that involve breaking down complex problems into sub-tasks and synthesizing solutions.

Should I use this for my use case?

This model is particularly well-suited for applications requiring:

  • Advanced Multimodal AI: If your application involves processing and generating content based on both text and images.
  • Long-Context Understanding: For tasks that demand comprehension and generation over extremely long documents, codebases, or conversational histories.
  • Complex Reasoning and Synthesis: Ideal for research or applications related to recursive task decomposition, planning, and synthesis, especially within the context of the Recursive Task Synthesis project. Its RL fine-tuning suggests improved performance in structured problem-solving.

Finetunes
1 models