Another admirer of DeepSeek has joined the ranks, such as Illia @ilblackdragon, co-founder of NEAR. In an interview, he stated that the team successfully tackled extremely complex code formal verification by equipping DeepSeek with a customized toolchain. When performing high-difficulty mathematical logic deductions, its rigor matches that of Claude 3.5 Sonnet and top-tier ChatGPT. Even more impressive: DeepSeek is 100 to 200 times cheaper than leading proprietary models—research tasks that once cost tens of thousands of dollars can now be completed for just a few hundred dollars, while completely avoiding the heavy-handed censorship of proprietary giants. It’s not that proprietary models are unaffordable—it’s just that open source offers far better value 😏.