Key points
- DeepSeek launches new AI model, DeepSeek-V4, with ultra-long context capability.
- Model released in two variants, targeting efficiency and high-performance applications.
- Announcement comes amid rising global competition and tensions in the AI sector.
Main story
DeepSeek has released its highly anticipated new artificial intelligence model, DeepSeek-V4, marking a significant development in the global AI race more than a year after its earlier breakthrough shook the industry.
The company announced on Friday that the new model features an “ultra-long context” capacity of up to one million words, allowing it to process and retain significantly larger volumes of information compared to earlier systems.
DeepSeek-V4 has been launched in two versions — V4-Pro and V4-Flash — with the former offering higher performance through 1.6 trillion parameters, while the latter provides a more cost-efficient alternative with 284 billion parameters.
According to the company, the model demonstrates strong capabilities across reasoning, world knowledge, and AI agent performance, positioning it among leading open-source systems globally. It is also optimised for integration with AI agent tools such as Claude Code, OpenClaw, OpenCode, and CodeBuddy.
A preview version of the open-source model is currently available to developers and researchers.
The issues
The launch underscores intensifying competition in the global AI industry, particularly between China and the United States. It also raises ongoing concerns about data privacy, transparency, and the governance of powerful AI systems.
What’s being said
DeepSeek claims its V4-Pro model rivals top-tier systems, including Gemini-Pro-3.1, while maintaining a cost-effective development approach.
The company previously disrupted the market with its R1 reasoning model, which delivered performance comparable to leading platforms such as ChatGPT but with significantly lower computing requirements — a development widely referred to as the “DeepSeek shock.”
Meanwhile, geopolitical tensions continue to shape the AI landscape. The United States has accused Chinese entities of attempting to extract proprietary AI technologies, signalling heightened scrutiny and potential regulatory action.
Chinese Premier Li Qiang has also highlighted the country’s leadership in advancing open-source AI ecosystems.
What’s next
DeepSeek is expected to expand deployment of its new model across sectors, building on its growing adoption in China’s public services, healthcare, and financial industries.
At the same time, global tech firms such as Meta and Microsoft are reportedly restructuring operations to prioritise AI investment, signalling continued acceleration in the sector.
Bottom line
DeepSeek’s latest AI model reinforces China’s growing influence in the global AI race, intensifying competition with Western tech giants and reshaping the future of open-source artificial intelligence.
