In a sense, yes. DeepSeek R1 was not the first to figure out the reasoning model tricks. We’ve already seen Google’s Gemini “thinking” model, Anthropic, and OpenAI’s o1 and o3 before DeepSeek R1, delivering similar or higher performances.
In a sense, yes. DeepSeek R1 was not the first to figure out the reasoning model tricks. We’ve already seen Google’s Gemini “thinking” model, Anthropic, and OpenAI’s o1 and o3 before DeepSeek R1, delivering similar or higher performances.