DeepSeek Poised to Unveil Latest AI Model
China’s DeepSeek is reportedly on the verge of releasing its latest artificial intelligence model.
That’s according to a report Friday (Feb. 27) by the Financial Times (FT), which noted that the debut of the company’s new large language model marks a fresh test of China’s efforts to challenge its American competitors in the AI space.
Two sources told the FT that DeepSeek plans to release V4, a “multimodal” model with picture, video and text-generating functions, and that DeepSeek had worked with Chinese AI chipmakers Huawei and Cambricon to optimize V4 for their newest releases.
According to the report, the launch is timed ahead of the yearly parliamentary “Two Sessions” meetings, which begin March 4. The FT says this high-profile political meeting could further establish DeepSeek as a national AI champion.
The report also pointed out that this will be DeepSeek’s first major release since the debut of its R1 reasoning model in January of last year. The company said that model could perform at the levels of top American models with just a fraction of the computing power.
That announcement shook the American tech world, with some experts comparing it to the Soviet Union’s launch of Sputnik in 1957.
DeepSeek could “catalyze a shift in Silicon Valley’s approach to AI,” Gokul Naidu, a consultant for SAP, told PYMNTS at the time. “Historically, the tech epicenter has prioritized growth at all costs, often ignoring questions of efficiency. But DeepSeek forces a recalibration. With its dramatically lower cost structure, businesses and developers alike are reevaluating their priorities, focusing on efficiency-driven growth.”
Since then, the FT added, DeepSeek has put out incremental updates rather than releasing new models, which has let Chinese rivals like Alibaba capture some of the demand for lower-cost, open-source AI models.
In other AI news, industry pioneer Andrew Ng said last week that the industry is decades away from developing artificial general intelligence (AGI).
He told Fast Company he defines AGI as AI that can perform the same intellectual feats that humans can, like learning to drive a truck or writing a Ph.D. thesis.
“We’re still very far from AI meeting that definition of AGI,” said. Ng, founder of DeepLearning.AI and Coursera, executive chairman of LandingAI, and founding lead of Google Brain.
He added that while some businesses have held up looser definitions of AGI, by his definition, “we are closer than before, yet many decades away from an AI that matches human intelligence. If you stick with the original definition — aligned with what people genuinely imagine AGI to be — we remain very, very far away.”
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
The post DeepSeek Poised to Unveil Latest AI Model appeared first on PYMNTS.com.