DeepSeek Challenges Claude Code! New AI Agent Team Launches Major Recruitment Drive

Chinese AI startup DeepSeek is publicly escalating its challenge to Anthropic's popular programming tool, Claude Code. The company has established a new AI agent team, launched official social media accounts, posted multiple job openings, and upgraded the agent capabilities of its flagship model, signaling a strategic shift toward AI agent products capable of autonomously executing complex tasks.

Headquartered in Hangzhou, DeepSeek recently opened a WeChat account named "DeepSeek Harness Team" on Tencent's super app. The account belongs to a Beijing-based company verified by Tencent. According to China's enterprise database Qichacha, this company is controlled by DeepSeek.

"Harness" typically refers to software frameworks built around large language models (LLMs) that enable AI agents to use tools, manage workflows, and complete more complex tasks. AI developers like Anthropic have used such frameworks to enhance their products, making them more efficient at handling programming and other professional tasks. Claude Code is currently one of the most prominent AI programming agent tools.

DeepSeek has also begun recruiting for multiple positions within the Harness team. In one job posting, the company stated its goal is to transform its models into state-of-the-art agent-based products, indicating a shift beyond focusing solely on foundational model performance and toward strengthening integration between models, tools, workflows, and real-world application scenarios.

Additionally, DeepSeek announced late Wednesday via its official WeChat chat channel that it has updated its flagship V4 Pro model, further improving AI agent capabilities. The near-simultaneous unveiling of the new team and model upgrade suggests DeepSeek is accelerating its efforts to build a complete product ecosystem spanning from underlying models to agent applications.

Chinese AI firms are currently engaged in fierce competition with U.S. developers to create more capable and cost-effective AI tools. Since OpenAI's ChatGPT sparked the global generative AI race, industry focus has gradually shifted from chatbots to AI agents—products capable of understanding instructions, operating software, and autonomously completing multi-step tasks—aiming to enter the enterprise automation market.

Compared to general chatbots, AI agents possess clearer commercialization potential, capable of helping programmers, researchers, and other professionals automate work, making them a key battleground in the new phase of competition between Chinese and U.S. AI firms. DeepSeek did not respond to requests for comment after regular business hours.