DeepSeek Launches New AI Models; Reuters: Market Response Lukewarm
Chinese AI startup DeepSeek unveiled new AI model preview versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, to a lukewarm market reception, a stark contrast to the widespread attention its previous low-cost models garnered. This indicates a shift in market expectations and increased competition from rivals like Kimi and Qwen, eroding DeepSeek's lead.
📋 Article Processing Timeline
- 📰 Published: April 27, 2026 at 19:17
- 🔍 Collected: April 27, 2026 at 19:31 (14 min after Published)
- 🤖 AI Analyzed: April 27, 2026 at 22:34 (3h 2m after Collected)
(Beijing, April 27, Central News Agency combined foreign report) Chinese artificial intelligence (AI) startup DeepSeek recently launched preview versions of its highly anticipated new AI models, but the market response has been lukewarm, a stark contrast to the situation over a year ago when it shook the industry with its low-cost AI models.
Reuters reported that over a year ago, DeepSeek released its DeepSeek-V3 and DeepSeek-R1 models, emphasizing that both products were trained using only a fraction of the computing resources of their American competitors. At the time, this garnered widespread international attention and led investors to question the rationality of investing heavily in AI infrastructure, triggering a sell-off in global tech stocks for a period.
However, after DeepSeek released its two new model versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, on the 24th, the market response was cold, highlighting a rapid shift in market expectations. The market and industry are increasingly accustomed to developing low-cost, high-performance models with limited computing resources, so the launch of the company's new models was not surprising.
Lian Jye Su, an analyst at market research firm Omdia, stated, "This launch largely follows a predictable trajectory," adding that academia and industry have widely explored advancements in model architecture and performance improvements in recent years.
AI analysis website Artificial Analysis pointed out that DeepSeek-V4 Pro shows significant improvement over previous versions but overall remains among the top open-source weighted models, not significantly surpassing competitors. Additionally, Chinese rivals such as Kimi and "Tongyi Qianwen" (Qwen) are rapidly catching up.
Su Lian Jye said, "Market expectations for the emergence of new competitors have already been reflected in valuations," emphasizing that the market is viewing the capabilities and limitations of AI more pragmatically.
The report noted that intensified competition in the Chinese market, with many players successively launching increasingly powerful models, is eroding DeepSeek's leading edge. (Compiler: Hong Qiyuan) 1150427
Choose to stand with facts, every sponsorship is a force to protect press freedom.
Download the Central News Agency's "First-hand News" APP to stay updated with the latest news.
The text, images, and audio/video on this website may not be reproduced, publicly broadcast, or publicly transmitted and used without authorization.
Reuters reported that over a year ago, DeepSeek released its DeepSeek-V3 and DeepSeek-R1 models, emphasizing that both products were trained using only a fraction of the computing resources of their American competitors. At the time, this garnered widespread international attention and led investors to question the rationality of investing heavily in AI infrastructure, triggering a sell-off in global tech stocks for a period.
However, after DeepSeek released its two new model versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, on the 24th, the market response was cold, highlighting a rapid shift in market expectations. The market and industry are increasingly accustomed to developing low-cost, high-performance models with limited computing resources, so the launch of the company's new models was not surprising.
Lian Jye Su, an analyst at market research firm Omdia, stated, "This launch largely follows a predictable trajectory," adding that academia and industry have widely explored advancements in model architecture and performance improvements in recent years.
AI analysis website Artificial Analysis pointed out that DeepSeek-V4 Pro shows significant improvement over previous versions but overall remains among the top open-source weighted models, not significantly surpassing competitors. Additionally, Chinese rivals such as Kimi and "Tongyi Qianwen" (Qwen) are rapidly catching up.
Su Lian Jye said, "Market expectations for the emergence of new competitors have already been reflected in valuations," emphasizing that the market is viewing the capabilities and limitations of AI more pragmatically.
The report noted that intensified competition in the Chinese market, with many players successively launching increasingly powerful models, is eroding DeepSeek's leading edge. (Compiler: Hong Qiyuan) 1150427
Choose to stand with facts, every sponsorship is a force to protect press freedom.
Download the Central News Agency's "First-hand News" APP to stay updated with the latest news.
The text, images, and audio/video on this website may not be reproduced, publicly broadcast, or publicly transmitted and used without authorization.