DeepSeek-V3.2-Speciale is a variant of the model that pushes the boundaries of reasoning capabilities, rivaling Gemini-3.0-Pro. It achieves gold-level results in IMO, CMO, ICPC World Finals, and IOI 2025, demonstrating its exceptional performance. However, it requires higher token usage and is currently available only through a temporary API endpoint, with no tool calls, until December 15th, 2025.
DeepSeek-V3.2 integrates thinking directly into tool-use, supporting both thinking and non-thinking modes. The model is designed to provide world-leading reasoning capabilities, making it an ideal choice for complex tasks. The open-source release of the model allows for community evaluation and research, and the tech report provides further details on its architecture and performance. The model's capabilities make it a valuable asset for various industries and applications.
