DeepSeek V4’s programming capabilities are “significantly ahead,” foreign media report that Chinese open-source AI is gaining global influence.

```

On April 24, DeepSeek officially released a preview version of its latest large language model V4, and plans to release it to the public as open source.

On the same day, The New York Times commented that “DeepSeek’s new work will further extend China’s global influence in the field of open-source artificial intelligence.” The article pointed out that Chinese companies have fully embraced the concept of “top-tier models open to all humanity”, and China’s open-source AI is becoming a consistently important global force.

The release of DeepSeek V4 is not only about technological iteration, but also marks that China’s AI industry has entered a new stage of normalized competition. Through technological openness and cost advantage, China is accelerating the narrowing of the technology gap with American giants like Anthropic and OpenAI, profoundly reshaping the global AI landscape.

This “extension of influence” has already produced substantial global effects. From Lagos to Kuala Lumpur, developers with limited budgets are flocking to Chinese models. In mid-2023, Malaysia’s deputy minister of communications publicly stated that the country’s sovereign AI infrastructure will be built on DeepSeek technology.

Kevin Xu, founder of the hedge fund Interconnected Capital, characterized this as “soft power in the technology field”: “The open-source strategy, by providing cheaper and more accessible tools, has effectively won recognition from the global developer community.”

“Far Ahead” in Programming Capability

DeepSeek V4 offers two versions, “Pro” and “Turbo”, targeting different application scenarios. The company claims that V4 excels at agent tasks, knowledge processing, and reasoning ability, and has been specially optimized for mainstream agent tools such as Anthropic’s Claude Code.

According to Vals AI’s evaluation, V4 is “far ahead” of all other open-source systems in code generation. Meanwhile, Moonshot AI recently released its latest open-source model Kimi 2.6. Vals AI CEO Rayan Krishnan commented that when compared horizontally with Moonshot AI’s latest Kimi 2.6, both perform about equally across various tasks.

It is worth noting that the release of V4 comes more than a year after DeepSeek’s previous flagship, R1. Released in January 2025, R1 achieved reasoning performance comparable to or even better than top large models at an extremely low computational cost, causing a stir in the global tech market and prompting a reassessment of the scale logic of AI infrastructure investment.

From the “DeepSeek Moment” to Normalized Competition

In January 2025, DeepSeek built a high-performance AI system at a chip cost far lower than that of OpenAI and Anthropic, shaking the entire industry. This event became known as China’s “DeepSeek Moment”—symbolizing Chinese AI companies’ ability to showcase their technical strength to the world.

Afterwards, Chinese companies kicked off a wave of intensive open-source model releases. According to The New York Times, citing data from the AI model marketplace platform OpenRouter, Chinese open-source models accounted for about a third of global AI usage in 2025, with DeepSeek ranked first and Alibaba closely following.

Alibaba has become a leader in this cohort. Its Qwen model series has accumulated over 1 billion downloads. TikTok’s parent company ByteDance, after investing $11 billion in AI infrastructure in 2024, also disclosed some technical details. Wei Sun, chief AI analyst at Counterpoint Research’s Beijing team, said that DeepSeek’s success paved the way for Chinese tech giants to publicly release AI systems rather than strictly keeping them sealed.

Kevin Xu, founder of the AI-focused hedge fund Interconnected Capital, said: “In 2025, this generation of open-source AI builders from China can be considered the most important story in the AI field. The speed and rhythm of model progress and releases, as well as the number of AI labs that compete and seemingly cheer each other on, are coming fast and hard, with no sign of slowing.”

Open-Source Strategy: China’s Global Soft Power in AI

China’s AI industry has fully bet on the open-source path, producing real global influence. From Lagos to Kuala Lumpur, developers with limited budgets are flocking to Chinese open-source models because of their lower operating costs and easier experimentation thresholds.

A recent research report by a US Congressional advisory agency pointed out that China’s push towards open-source AI has formed a significant economic advantage domestically—because of extremely low usage thresholds, the related systems quickly penetrate into robotics, logistics, manufacturing and other industrial fields, and real-world data generated from industrial applications feed back into the continuous iteration of AI systems.

Kevin Xu characterized this trend as a logic of soft power in the technology field: “Open source is the soft power of the future technology field.” He noted that as Chinese companies expand overseas, the open-source strategy, by providing cheaper and more accessible tools, has effectively won recognition from the global developer community.

DeepSeek’s choice of open source is also part of China’s overall AI strategy. Analysts said that China has almost fully opened its top systems in open-source models to the outside world, forming a sharp contrast to the tight protection of core models by American companies like OpenAI and Anthropic.

Risk Warning and DisclaimerThe market is risky; investment needs caution. This article does not constitute personal investment advice, nor does it take into account the specific investment goals, financial circumstances, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article fit their specific situation. Any investment based on this is at your own risk. ```