人工智能动态跟踪：计算机行业：DeepSeek提振国产大模型信心，有望促进AI行业发展

Investment Rating - The report maintains a "Positive" investment rating for the computer industry, indicating an expectation of performance exceeding the market benchmark by more than 5% [6]. Core Insights - The launch of DeepSeek-V3 and R1 has boosted confidence in domestic large models and is expected to accelerate the commercialization of AI products [4]. - DeepSeek-V3, with 671 billion parameters, has shown superior performance compared to other open-source models and is competitive with top closed-source models like GPT-4o [2]. - The introduction of DeepSeek-R1, utilizing reinforcement learning, significantly enhances inference capabilities with minimal labeled data, achieving performance comparable to OpenAI's o1 version [2]. - The efficiency of DeepSeek-V3 has improved, with generation speed increasing from 20 TPS to 60 TPS, marking a threefold enhancement [2]. - The cost of using DeepSeek models has been significantly reduced, with API pricing adjustments making it more accessible for users [2]. Summary by Sections Model Performance and Innovation - DeepSeek-V3 outperforms Qwen2.5-72B and Llama-3.1-405B in multiple evaluations, matching the performance of leading models like GPT-4o and Claude-3.5-Sonnet [2]. - The model's inference efficiency has tripled, enhancing user experience [2]. Market Impact and User Engagement - DeepSeek applications achieved 2.59 million daily active users within 15 days, surpassing ChatGPT's user base, indicating strong market demand for enhanced model capabilities [3]. - The cost reduction in large models is expected to drive user growth and accelerate the commercialization process across the AI industry [3]. Investment Recommendations - Companies likely to benefit from the developments include infrastructure suppliers like Zhongke Shuguang and Inspur Information, as well as firms actively exploring AI applications such as iFLYTEK and Kunlun Wanwei [4].