Gemini 2.0 Flash Thinking
Early experimental reasoning variant of Gemini 2.0 Flash that shows its chain of thought
Gemini 2.0 Flash Thinking is an experimental reasoning variant of Gemini 2.0 Flash that Google released on December 19, 2024, shortly after the base 2.0 Flash preview. Unlike a standard chat model, it generates and shows an intermediate chain of thought before giving its final answer, similar in spirit to OpenAI’s o1, which lets users see the steps the model worked through on math, logic, and coding problems. It kept the million-token context window and multimodal input of the base Flash model.
Google shipped it first through Google AI Studio and the Gemini API as an experimental preview, then extended it into the Gemini app and connected it to Search, Maps, and YouTube by February 2025. It served as Google’s early public test of reasoning-focused prompting techniques before the company folded thinking capabilities directly into the main Gemini 2.5 line, which made the standalone Flash Thinking variant unnecessary going forward.