Source of information:

https://gizmodo.com/deepseek-v-3-1-release-aider-score-china-ai-2000645531

The V3.1 model released by the Chinese AI company DeepSeek is shaking up the global AI competition landscape. Its ultra-low cost of just $1.01 per coding task directly challenges the pricing systems of American tech giants like OpenAI and Anthropic, which often start at $70. This 685 billion parameter model achieved a high score of 71.6% on the Aider coding benchmark test, surpassing Claude Opus 4's 70.6%, yet its cost is only 1/68th of the latter's. Industry analysts believe this not only marks a new phase in Sino-US AI competition but could also trigger a fundamental shift in the entire industry from pursuing performance to emphasizing cost-effectiveness.

The significance of this breakthrough goes beyond mere technological advancement. As a relatively small Chinese company, DeepSeek has gained such a significant advantage in a direct confrontation with well-funded American tech giants, showcasing the rapid rise of China's AI innovation ecosystem. More importantly, its cost advantage may completely change the speed and scope of AI technology adoption, making high-end AI services once only affordable for large enterprises accessible to all.

Strategic Intentions Behind the Technological Breakthrough

Elon Musk and Sam Altman of OpenAI. © Muhammed Selim Korkutata/Anadolu via Getty Images

The technical architecture of DeepSeek V3.1 reflects a different development philosophy compared to its American counterparts. The most remarkable feature of this model is its extended context window, capable of processing larger amounts of data than traditional models. This design allows the model to maintain coherence during long conversations, use memory to complete complex repetitive tasks, and better understand the relationships between different parts of the text.

From a technical implementation perspective, a larger context window means users can handle longer documents and engage in deeper conversations in a single session without frequently restarting or repeating background information. This capability holds significant value for tasks such as programming, document analysis, and creative writing that require continuous context.

AI researcher Andrew Christianson commented on social media: "DeepSeek V3.1 scored 71.6% on the Aider benchmark test, 1% higher than Claude Opus 4, but it is 68 times cheaper." This comparison data clearly highlights DeepSeek's overwhelming advantage in cost-effectiveness. Notably, the model immediately caused a sensation upon its release on the renowned AI tool testing platform Hugging Face, quickly surpassing competitors like OpenAI's ChatGPT 4.5.

DeepSeek's success was no accident. The R1 model previously released by the company had already shocked AI observers with its speed and accuracy, showing a clear advantage over Western competitors. The release of V3.1 further solidifies DeepSeek's position in the AI field and proves the rapid progress of Chinese companies in this key technology sector.

Re-defining the Rules of AI Competition

The emergence of DeepSeek V3.1 may mark a fundamental shift in the center of AI competition. Traditionally, American tech giants have focused on building the most powerful and advanced models, often investing resources without considering costs to pursue performance limits. While this strategy has achieved significant results in technical indicators, it has also led to high usage costs, limiting the reach of AI technology.

In contrast, DeepSeek adopts a more pragmatic approach: significantly reducing usage costs while ensuring performance levels. The success of this strategy indicates that AI competition may be shifting from a purely technical arms race to a focus on practicality and accessibility.

Louis Liang, an AI investor at Ameba Capital, told Bloomberg: "Although many people recognize DeepSeek's achievements, this is just the beginning of the wave of Chinese AI innovation. We are witnessing the arrival of an era of large-scale AI applications, which goes beyond national competition." This viewpoint reveals the deeper implications of the current transformation: AI technology is moving from laboratories to the mass market, and cost-effectiveness will become the key factor in determining success.

VentureBeat reported that DeepSeek's approach may shift global competition from focusing on functionality to emphasizing accessibility. If this trend becomes widespread, it will have a profound impact on the entire tech industry. Tech giants like Meta and Alphabet may need to re-evaluate their AI strategies, seeking ways to reduce costs while maintaining technological leadership.

New Landscape of the Global AI Ecosystem

DeepSeek's rise is not an isolated phenomenon but a reflection of the overall strength of China's AI industry. Companies such as Alibaba Group and Moonshot AI have also released AI models challenging American tech companies, creating a competitive situation with multiple points of success.

This diversified competitive landscape has positive implications for global AI development. First, it breaks the monopoly of a few American companies on high-end AI technology, offering more choices to the market. Second, intense competition drives technological innovation and cost optimization, ultimately benefiting global users.

From a geopolitical perspective, the competition between China and the US in the AI field reflects the struggle for technological dominance between the two countries. However, as Louis Liang pointed out, the significance of this competition has gone beyond the national level, focusing more on how AI technology can better serve human society.

The success of DeepSeek V3.1 also demonstrates the unique wisdom of Chinese AI companies in choosing technical paths. Compared to American companies often pursuing technological extremes, Chinese companies emphasize the practicality and accessibility of technology. This differentiated development strategy not only helps Chinese companies find their own position in the fierce competition but also provides new ideas for global AI development.

Currently, the DeepSeek team is planning to use petawatt lasers to further verify and improve its technical solutions in real-world applications. As the technology continues to mature and the application scenarios expand, this low-cost, high-performance AI model is expected to promote the application of artificial intelligence in more fields, truly achieving the democratization of AI technology.

For the entire AI industry, the emergence of DeepSeek V3.1 may mark the beginning of a new era: an era that places greater emphasis on practicality, accessibility, and cost-effectiveness. In this era, the value of technology is no longer solely reflected in the scale of parameters and benchmark test scores, but in its ability to provide truly valuable services to more users. This transformation holds profound significance for promoting the global spread and application of AI technology.

Original article: https://www.toutiao.com/article/7540965215400247858/

Disclaimer: The article represents the views of the author. Please express your opinion below using the [Up/Down] buttons.