DeepSeek launches new AI model with long context

ISLAMABAD: (Web Desk) – Chinese AI startup DeepSeek has unveiled preview versions of its latest artificial intelligence model, DeepSeek-V4, marking a significant step forward more than a year after its earlier breakthrough in low-cost AI development.

According to the company, the new model features an ultra-long context window of up to one million words, positioning it among the most advanced open-source AI systems in terms of reasoning ability, world knowledge, and agent-based capabilities.

DeepSeek-V4 has been introduced in two variants — DeepSeek-V4-Pro and DeepSeek-V4-Flash. The Flash version is designed to be more efficient and cost-effective, using smaller parameters while maintaining strong performance.

The company claims that DeepSeek-V4-Pro performs at the top level among open-source models and comes close to leading proprietary systems such as Gemini-Pro-3.1 developed by Google.

Deepika Padukone Ranveer Singh Expect Second Child

The release is expected to intensify global competition in the AI sector, where companies are racing to build more powerful and scalable models capable of handling increasingly complex tasks.