appConfigEntityId

bodyType

customContents

entityId

entityUuid

isVideo

articlePageQuery

article

articleSchemaContent

content

hooksArticlePageAdvertisementConfigArticle

hooksArticlePageFloatingButtonConfigArticle

appConfig

filter

json

liveArticle

Xinmei Shen

DeepSeek announced on Thursday a “significant” price hike, amid surging global demand for a new ultra-cheap model, which sparked debate over how the Chinese artificial intelligence start-up has delivered powerful performance at rock-bottom prices.
Overall pricing for its application programming interface (API) services would be raised “in the near future”, with specific plans to be advised later, the Hangzhou-based company said in a statement on its developer platform on Thursday.
A “significant increase” was expected, and users should plan their usage accordingly, DeepSeek added.
The notice came a week after the release of its latest model DeepSeek-V4-Flash-0731, a 284-billion-parameter lightweight version of the V4 series previewed in April.
The company’s plans for a major price hike underscore the challenges the Chinese start-up faces to maintain its aggressively low prices amid fierce competition, potentially clouding its reputation for being a champion for cost-effective AI.
“DeepSeek choosing to raise prices at this time – isn’t this just asking for trouble?” wrote AI developer Michael Guo in a social media post on Thursday, arguing that new American models including Meta’s Muse Spark and OpenAI’s GPT-5.6 Luna were competitive with DeepSeek models in both capabilities and price.
DeepSeek’s new model rapidly gained traction in the global AI community for combining near-frontier capabilities with extremely low prices.


DeepSeek-V4-Flash-0731 was currently the second strongest open-weight AI model on the market, behind only the massive 2.8-trillion-parameter model Kimi K3 from Beijing-based Moonshot AI, and came “roughly midway” between US lab Anthropic’s Claude Opus 4.5 and Opus 4.6, research firm Epoch AI said on Wednesday.
For now, DeepSeek-V4-Flash-0731 on average costs 3 US cents per task, making it roughly 105 times cheaper than Anthropic’s Claude Fable 5, which costs US$3.15 per task, benchmark firm Artificial Analysis said earlier this week.
What was notable about DeepSeek’s latest model was how cheap it was even compared with its Chinese open-weight peers, Epoch AI said, noting that the current price at 14 US cents per million input tokens and 28 US cents per million output tokens was an “order of magnitude cheaper” than Zhipu AI’s GLM-5.2 first-party API, which costs US$1.40 per million input tokens and US$4.40 for output.
DeepSeek-V4-Flash-0731 became the “fastest growing model ever” in terms of token usage on Ollama, a platform that helps users run AI models on local hardware, the Palo Alto, California-based company said on social media on Wednesday, adding it was expanding capacity for the model in the US and Europe.
The Chinese start-up was able to offer such competitive pricing thanks to both architectural optimisation and an efficient training strategy, said Zhenhui Jack Jiang, professor of innovation and information management at the University of Hong Kong’s Business School.
Its mixture-of-experts design, for instance, activated only a small subset of experts for each input, allowing it to maintain substantial capacity while using much less computation and memory during inference, Jiang said.
DeepSeek’s training method of multi-teacher knowledge distillation, which combined the capabilities of several specialised models into one unified model, also allowed it to absorb useful capabilities more efficiently, he added.


Top Chinese AI models make 10% of OpenAI, Anthropic revenue despite high valuations: report

China to see major shift to Huawei for AI model training in 2027: deputy chair

People’s Daily rejects US claims of malicious AI distillation, warns of countermeasures

DeepSeek AI engineer slams Anthropic, OpenAI over ‘pacing’ calls, invokes Nazi Germany

DeepSeek

DeepSeek signals ‘significant’ price hike amid surge in demand for low-cost AI models

A notice of price increase comes a week after the release of DeepSeek's latest model, a 284-billion-parameter lightweight version of the V4 series. Photo: Reuters

OpenAI blinks in face-off with Chinese rivals, drops pricing for some models up to 80%

OpenAI CEO Sam Altman arrives for a meeting at the White House in Washington on July 30. Photo: Reuters

People visit the Moonshot AI booth during the World Artificial Intelligence Conference in Shanghai on July 18. Photo: AFP

China’s DeepSeek beefs up agentic AI with ‘harness’ tests as V4 model jolts Silicon Valley

A smartphone screen displays the logo for Chinese AI company DeepSeek. Illustration: Getty Images

China’s AI revenue projected to reach US$13b on breakthroughs, adoption: Goldman Sachs

Goldman Sachs boosted its year-end annualised recurring revenue projection for mainland Chinese AI models to US$13 billion from US$10 billion. Photo: Shutterstock

The MiniMax booth at the World AI Conference in Shanghai, China, on July 17. Photo: AP

Tencent ramps up overseas AI push with wider roll-out of flagship Hy3 model

The Tencent logo. Photo: Reuters

The Tencent booth during the World Artificial Intelligence Conference in Shanghai on July 18. Photo: AFP

Tech

Tech Trends

News

China Future Tech

Plans for a major price hike underscore the challenges the company faces to maintain aggressively low prices amid fierce competition


Plans for a major price hike underscore the challenges the company faces to maintain aggressively low prices amid fierce competition.


DeepSeek signals ‘significant’ price hike amid surge in demand for low-cost AI models

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}Plans for a major price hike underscore the challenges the company faces to maintain aggressively low prices amid fierce competition

Plans for a major price hike underscore the challenges the company faces to maintain aggressively low prices amid fierce competition