No products in the cart.
Introduction
Chinese AI lab DeepSeek, which recently launched DeepSeek-V3, is back with yet another powerful reasoning large language model named DeepSeek-R1. The new model has the similar mixture-of-experts architecture and matches the performance of OpenAI’s frontier model o1 in tasks like math, coding and general knowledge.
The DeepSeek-R1 is reportedly 90–95 per cent more affordable than o1.💸
What is DeepSeek-R1?
The new AI model from DeepSeek is a state-of-the-art reasoning model designed to enhance problem solving and analytical capabilities of AI systems. Based on the research paper, the new model comprises two core versions — DeepSeek-R1-Zero and DeepSeek-R1.
GitHub – deepseek-ai/DeepSeek-R1
Why is DeepSeek R1 special?
DeepSeek has unveiled its latest model, DeepSeek-R1, marking a significant stride toward advancing artificial general intelligence (AGI) — AI capable of performing intellectual tasks on par with humans. As teams increasingly focus on enhancing models’ reasoning abilities, DeepSeek-R1 represents a continuation of efforts to refine AI’s capacity for complex problem-solving.
OpenAI had previously set a benchmark in this domain with its o1 model, which leverages chain-of-thought reasoning to break down and solve problems step by step. Using reinforcement learning (RL), o1 improves its reasoning strategies by optimizing for reward-driven outcomes, enabling it to identify and correct errors or explore alternative approaches when existing ones fall short.
Building on this foundation, DeepSeek-R1 employs a hybrid approach that combines reinforcement learning with supervised fine-tuning to tackle challenging reasoning tasks. The model has demonstrated competitive performance, achieving 79.8% on the AIME 2024 mathematics tests, 97.3% on the MATH-500 benchmark, and a 2,029 rating on Codeforces — outperforming 96.3% of human programmers. For comparison, OpenAI’s o1–1217 scored 79.2% on AIME, 96.4% on MATH-500, and 96.6% on Codeforces.
In terms of general knowledge, DeepSeek-R1 achieved a 90.8% accuracy on the MMLU benchmark, closely trailing o1’s 91.8%. These results underscore DeepSeek-R1’s capability to handle a broad range of intellectual tasks while pushing the boundaries of reasoning in AGI development.
Let’s see how to use DeepSeek R1 locally
LM Studio – Discover, download, and run local LLMs
Run Llama, Mistral, Phi-3 locally on your computer.
2. After installation , go to the Discover Tab🔍
Select any one of the two , and download the model from the right hand side panel.
Click Download.
3. Once you click download the model will be loaded in the memory:
and from here you can ask questions from DeepSeek R1.
Far more affordable than o1
In addition to enhanced performance that nearly matches OpenAI’s o1 across benchmarks, the new DeepSeek-R1 is also very affordable. Specifically, where OpenAI o1 costs $15 per million input tokens and $60 per million output tokens, DeepSeek Reasoner, which is based on the R1 model, costs $0.55 per million input and $2.19 per million output tokens.
The model can be tested as “DeepThink” on the DeepSeek chat platform, which is similar to ChatGPT. Interested users can access the model weights and code repository via Hugging Face, under an MIT license, or can go with the API for direct integration.
Conclusion
In conclusion, DeepSeek R1 is a groundbreaking AI model that combines advanced reasoning capabilities with an open-source framework, making it accessible for both personal and commercial use. Its unique architecture allows for efficient computation while achieving impressive accuracy in complex tasks. The model’s focus on logical inference sets it apart from traditional language models, fostering transparency and trust in its outputs. With competitive pricing and local deployment options, DeepSeek R1 democratizes access to powerful AI tools. Overall, it represents a significant step forward in the evolution of reasoning-focused artificial intelligence.
—
Mouin Frada, Innovation Engineer
