DeepSeek has posted job listings for a new team focused on competing in this market for AI agents that help with more complex tasks such as coding. One posting said the company is looking to develop cutting-edge agentic products, according to the report.
The company announced Wednesday that it updated its flagship model, V4 Pro, to include better agentic capabilities, the report said.
DeepSeek also created an official social media account on the Chinese super app WeChat for the new “DeepSeek Harness Team.” The term “harness” refers to a component of software that helps AI agents perform complex tasks, per the report.
PYMNTS reported Aug. 5 that several AI companies have been working to enhance their offerings in AI coding.
For example, Meta announced Aug. 5 that it released a terminal coding agent called Muse Code in beta; OpenAI announced in June that that it plans to expand the capabilities of its AI coding agent, Codex, by acquiring Ona, a provider of secure cloud execution and orchestration technology; and it was reported Aug. 5 that Google is in talks with AI coding startup Mechanize to hire some of Mechanize’s talent and sign a non-exclusive licensing agreement for the company’s technology.
It was reported Thursday (Aug. 6) that DeepSeek has resumed its second funding round and is looking to raise close to $8 billion at a valuation of about $74 billion.
The report said the company is “riding high” after releasing its V4 Flash AI model, which gained attention for its cost-performance offering.
DeepSeek wrapped its first round of funding near the end of May, raising $7 billion at a $52 billion valuation.
The company’s unusually fast funding schedule was related to expectations of increased capital needed to construct its own data center and acquire more chips.
DeepSeek initially gained attention in January 2025 when it debuted an AI model that sent shockwaves through the AI world by offering performance comparable to those of American rivals OpenAI and Meta while using substantially fewer Nvidia chips.
