SL-AI/GRaPE-2-Pro

VISIONConcurrency Cost:2Model Size:27BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Apr 19, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

SL-AI/GRaPE-2-Pro is a 27 billion parameter multimodal language model developed by Skinnertopia Lab for Artificial Intelligence (SLAI), built on a Qwen3.5 base architecture. It accepts image and text inputs to produce text outputs, featuring an extended thinking mode system for controllable reasoning depth. The model is post-trained with a heavy emphasis on code, STEAM, logical reasoning, and structured problem solving, making it suitable for complex analytical tasks.

Loading preview...

GRaPE 2 Pro: Flagship Multimodal Reasoning Model

GRaPE 2 Pro is the 27 billion parameter flagship model from Skinnertopia Lab for Artificial Intelligence (SLAI)'s second-generation GRaPE family. Built on a robust Qwen3.5 base, it supports multimodal inputs (image + text) and generates text outputs. This model significantly improves upon its predecessor by incorporating a stronger base model, expanded thinking modes, and high-quality, proprietary training data.

Key Capabilities & Features

  • Multimodal Input: Processes both images and text to generate comprehensive text responses.
  • Controllable Reasoning Depth: Features six discrete <thinking_mode> tiers (minimal, low, medium, high, xtra-Hi, auto) allowing users to specify or let the model adapt its reasoning intensity for tasks.
  • Specialized Training: Post-trained on a curated proprietary dataset with a strong focus on code (~50% of data), STEAM (Science, Technology, Engineering, Arts, and Mathematics), logical reasoning, and structured problem-solving.
  • Enhanced Performance: Leverages a Qwen3.5-27 base and more parameters (27B) to deliver strong performance on structured reasoning tasks.

Ideal Use Cases

GRaPE 2 Pro is particularly well-suited for applications requiring deep analytical work, complex coding tasks, and multi-step mathematical problem-solving. Its controllable thinking modes allow for efficient resource allocation, making it adaptable for both quick queries and intensive computational reasoning.