DeepSeek, a Chinese AI firm, recently released an open version of its reasoning model, R1. The model reportedly matches or outperforms OpenAI’s o1 on some benchmarks.

DeepSeek has claimed that the cost to train its model was US$5.6 million, significantly lower than the hundreds of millions spent by some leading US companies.

Despite facing US export restrictions on essential chips, DeepSeek claims to have developed innovative strategies to address these challenges.

The announcement has sparked mixed reactions.

Venture capitalist Marc Andreessen acknowledged DeepSeek’s advancements, while journalist Holger Zschaepitz questions its economic impact on large-scale AI investments.

Y Combinator CEO Garry Tan suggested decreasing model development costs could accelerated industry growth.

Yann LeCun, Meta’s chief AI scientist, emphasized the importance of open-source innovation and collaboration in AI development.

Recent DeepSeek developments