🔥DeepSeek Surpasses ChatGPT, Tops Free App Download Chart in the U.S. App Store

#DeepSeek app has topped the free app download charts in both the China and U.S. App Stores, surpassing #ChatGPT in the U.S. download rankings.

According to reports, DeepSeek released the open-source DeepSeek-R1 last week, claiming that the model's performance is on par with OpenAI's o1 official version, which has led to heated discussions in the overseas AI industry regarding this model. The focal point of the debate is how the capabilities of open-source models catching up with the latest closed-source models could potentially change the competitive landscape of large models.

DeepSeek stated that DeepSeek-R1 scored closely to the o1 official version in tests such as Codeforces, GPQA Diamond, MATH-500, MMLU, and SWE-bench Verified, with some tests showing scores that even exceed those of the o1 official version. The model significantly enhanced its inference capabilities through large-scale reinforcement learning techniques during the post-training phase, despite having only a minimal amount of labeled data.

AI industry insiders are already discussing the implications of open-source software. Yann Lecun, Chief AI Scientist at Meta, stated that the emergence of DeepSeek-R1 signifies not so much that Chinese companies are surpassing American companies in the AI field, but rather that open-source large models are surpassing closed-source ones. "DeepSeek will benefit from open research and open source, similar to Meta's PyTorch and Llama. They propose new ideas and build upon the work of others. Because their work is public and open source, everyone can benefit from it, which is the power of open research and open source," Lecun remarked.

Marc Andreessen, founder of Silicon Valley venture capital firm A16Z, also commented that DeepSeek-R1 is one of the most stunning and impressive breakthroughs he has ever seen, and as an open-source model, its emergence brings a gift to the world. (Shangguan News)