Skip to main content

Exclusive news, data and analytics for financial market professionalsLearn more aboutRefinitiv

Deepseek AI sign in Beijing

Item 1 of 2 A DeepSeek AI sign is seen at a building where the Chinese start-up's office is located in Beijing, China, February 19, 2025. REUTERS/Florence Lo/File Photo

[1/2] A DeepSeek AI sign is seen at a building where the Chinese start-up's office is located in Beijing, China, February 19, 2025. REUTERS/Florence Lo/File Photo Purchase Licensing Rights, opens new tab

  • Summary

  • Companies

  • DeepSeek looking to regain momentum after being surpassed by Chinese rivals

  • V4-Flash about 105 times cheaper than Anthropic's Claude Fable 5, tests show

  • On performance, ranks on par with Google's Gemini 3.6 Flash but behind OpenAI and Anthropic ​offerings

BEIJING, Aug 3 (Reuters) - A version of Chinese startup DeepSeek's flagship AI model is by ‌far the least expensive to run on benchmark tests among well-known models globally and more than 100 times cheaper to run than Anthropic's Claude Fable 5, according to a research firm.

DeepSeek, which sources have said is preparing for a potential IPO, officially released its V4-Flash ​model on Friday, its latest attempt to regain momentum by doing what it is best known for - ​offering ultra-low-cost AI alternatives.

Sign up here.

The startup's R1 model became a global sensation in early 2025, triggering ⁠a selloff in global technology stocks and raising questions about the large amounts U.S. companies were spending on ​AI.

DeepSeek's V4-Flash charges $0.14 per million input tokens and $0.28 per million output tokens, according to research firm Artificial Analysis. A ​token is a unit of data used to measure AI usage.

San Francisco-based Artificial Analysis estimated V4-Flash's average cost at 3 cents per test, compared with 86 cents for Kimi K3 from Chinese rival Moonshot AI, $1.86 for OpenAI's GPT-5.6 Sol and $3.15 for Claude Fable 5.

The ​comparison provides a more realistic measure of value than pricing alone because it accounts for the amount of ​data a model must process and generate to complete a task. A model with low headline price can still prove expensive ‌if it ⁠requires significantly more steps to produce an answer.

DeepSeek once commanded most of the headlines about Chinese AI development but was quickly besieged by many domestic rivals including other startups such as Moonshot, MiniMax and Z.AI as well as tech giants like ByteDance and Alibaba (9988.HK), opens new tab. All are vying with U.S. tech firms for global adoption, targeting businesses seeking cheaper ​ways to deploy AI ​at scale.

Artificial Analysis said DeepSeek's ⁠V4-Flash model scored 50 out of 100 on its Intelligence Index, which combines results from nine benchmarks spanning coding, reasoning and workplace-style assignments.

That's the same score as Google's (GOOGL.O), opens new tab ​Gemini 3.6 Flash, and one point behind Meta's (META.O), opens new tab Muse Spark 1.1 and GLM-5.2 ​from Z.AI which ⁠is also known as Zhipu.

Moonshot's Kimi K3, however, scored a 57 while Anthropic's Claude Opus 5, Fable 5, and OpenAI GPT-5.6 scored nine or more points higher.

DeepSeek is also preparing a more powerful version of its model, called the ⁠V4-Pro. It ​has not given a date for that version's official release.

Separately on ​Monday, Alibaba unveiled its largest and most capable artificial-intelligence model to date, the Qwen3.8-Max, which is not far behind in size when compared with an ​offering from domestic rival Moonshot AI launched last month.

Reporting by Eduardo Baptista; Editing by Miyoung Kim and Edwina Gibbs

Our Standards: The Thomson Reuters Trust Principles., opens new tab

  • X

  • Facebook

  • Linkedin

  • Email

  • Link

Purchase Licensing Rights

Eduardo Baptista

Thomson Reuters

Eduardo Baptista is a Senior Correspondent for Reuters based in Beijing, covering China’s technology, space, and automotive industries. He has led enterprise and investigative reporting on China’s military-linked companies, artificial intelligence and semiconductor supply chains, as well as macroeconomic and industrial policy. Baptista has reported from China for nearly a decade and holds a BA in History from the University of Cambridge.

Read Next

  • ago

Alibaba unveils its most capable AI model to date, not far behind Moonshot's in size

EXCLUSIVE

Chinese military researchers tap US AI models to train defence systems

EXCLUSIVE

OpenAI discovered other autonomous agent escapes in hacking probe

Alphabet rolls back AI image generation in Google Earth over policy violations

Will SpaceX paint its rocket pink? Investor questions go beyond Moon and Mars ahead of first results

With Situational Awareness AI deal, Citadel's Griffin rides to the rescue again

Business

Businesscategory · August 3, 2026 · 5:37 AM UTC · ago

Italian cable maker Prysmian said on Monday it agreed to acquire U.S.-based electrical products ​maker Atkore for $95 per share in cash, giving ‌the company an implied enterprise value of around $3.8 billion.

5:35 AM UTC

5:32 AM UTC

5:20 AM UTC

5:14 AM UTC

Read Original at Reuters →