DeepSeek R1 is DeepSeek's open-weight reasoning model released in January 2025, offering capability in the same class as OpenAI o1. Its aggressive training-cost and pricing posture triggered a market shock the industry now calls the DeepSeek Moment. By publishing both a technical report and the weights it proved the Reasoning Models category isn't owned exclusively by closed labs. It is a clear inflection point that shifted the Open Source LLM balance.
MEVZU N°124ISTANBULYEAR I — VOL. III
Glossary · Beginner · 2025
DeepSeek R1
DeepSeek's 2025 open-weight reasoning model, which shook the industry with low-cost long-form thinking.
- EN — English term
- DeepSeek R1
- TR — Turkish term
- DeepSeek R1