Loading timeline…
20252020s
DeepSeek-R1 Release
DeepSeek released R1 on January 20, 2025, an open-weight reasoning model rivaling OpenAI's o1 at a fraction of the training cost.
Why It Mattered
Trained with large-scale reinforcement learning on minimal annotated data, R1 matched OpenAI's o1 on mathematics, coding, and reasoning benchmarks while being freely available as open weights. Its release triggered a global market shock — including a historic single-day selloff in AI-related stocks — by demonstrating that frontier reasoning capability did not require frontier-scale budgets, and it reset assumptions about China's position in the AI race.
Who Built It
DeepSeek
Applications
- Reasoning
- Open-source AI
- Mathematics