DeepSeek V4 is about to be released, with programming capabilities aiming to surpass Claude and GPT. Will Silicon Valley shake again next month?

According to the latest news, DeepSeek will release its next-generation flagship AI model V4 in February. According to The Information, this new model possesses powerful programming capabilities, with preliminary internal tests indicating it has surpassed Anthropic’s Claude and OpenAI’s GPT series. This is only two months after the previous V3 model (released in December 2024), demonstrating DeepSeek’s accelerated iteration speed once again. Considering the significant impact DeepSeek made when it emerged at the beginning of last year on Silicon Valley and the global AI market, the release of V4 may once again trigger market waves.

V4 Programming Capabilities Become a Core Selling Point

Supported by Internal Test Data

According to information from two insiders cited by The Information, DeepSeek V4’s performance on programming tasks has already exceeded that of the top models currently on the market. This is particularly noteworthy because programming ability has always been an important indicator of large model performance. Programming involves complex logical reasoning, code generation, debugging, and other dimensions, serving as a comprehensive test of a model’s understanding and execution accuracy.

Benchmarking Against Competitors

Based on relevant information, the main programming capability benchmarks in the current market include:

According to data, Grok 4’s “Battle of the AIs” ranking has risen by 8.2%, leading DeepSeek V3’s 5.9%, but the launch of V4 could change this landscape.

Release Date Approaching, Market Expectations High

Specific Release Plan

DeepSeek plans to release V4 around the Chinese New Year of 2025, which falls in February (Lunar New Year), only 3-4 weeks from now. However, according to the latest information, the release schedule may be adjusted based on actual circumstances, meaning the exact release date is subject to change.

Market Enthusiasm

From the performance of the A-share market, it is evident that investors are highly attentive to DeepSeek’s new developments. According to data, on the January 9th Dragon and Tiger list, DeepSeek concept stocks such as Kunlun Wanwei and Zhongcheng Technology received significant capital attention, with retail investors deploying over 5 billion yuan in total. This reflects that the market’s anticipation for V4’s release has already begun to be priced in.

DeepSeek’s Position Continues to Strengthen

From Market Recognition to Technical Leadership

According to the 2025 Fortune China Top 50 Tech Companies list, DeepSeek ranks second, only behind Huawei. This is extremely rare for a startup founded less than three years ago. More notably, traditional internet giants Tencent, Alibaba, and ByteDance did not make the top three, indicating that general AI technology has become a core aspect of China’s technological competitiveness.

From V3 to V4 Iteration Logic

After the release of V3 in December 2024, it quickly gained market attention due to its low cost and strong performance. The launch of V4 continues this trajectory, further enhancing capabilities in core application scenarios such as programming. Behind this rapid iteration is a reflection of DeepSeek’s strong investment in R&D and keen market demand awareness.

Key Areas to Watch Moving Forward

• Specific Release Date: Although scheduled around the Chinese New Year in February, the exact timing still requires official confirmation.
• Performance Data Release: Internal tests surpass Claude and GPT, but public third-party evaluations are needed.
• Application Ecosystem Response: The adoption speed of related app developers and enterprise clients after V4’s release.
• Market Reaction Intensity: Referencing the impact of V3 on the global AI market, V4 may bring a new wave of market volatility.
• Regulatory Developments: China’s stance on AI technology exports may influence V4’s international promotion.

Summary

The upcoming release of DeepSeek V4 once again puts this Chinese AI startup in the spotlight. Its competitive edge in programming capabilities, the imminent launch date, and high market expectations together suggest that February could become another pivotal moment in the global AI market. From the emergence of DeepSeek V1 at the beginning of last year to the anticipation of V4 now, this company has demonstrated China’s progress in general AI through concrete actions. The key next step is whether V4 can deliver on its promise of “leading programming capabilities” in practical applications, and whether this leadership can translate into market share and ecosystem influence. For the entire AI industry, intensified competition ultimately benefits users.