DeepSeek drops enhanced coding agent with aim to rival Anthropic

DeepSeek has unveiled an open-source autonomous coding agent that can monitor developer discussions and suggest solutions, intensifying its competition with Anthropic's Claude Code.

DeepSeek has released a prototype of a more autonomous coding agent as it seeks to challenge Anthropic PBC’s popular Claude Code. 

 The new open-source offering, dubbed “Super-agent” can proactively monitor group chats among developers for feedback on coding issues and suggest solutions without being summoned, the DeepSeek Harness team said in a post on WeChat on Saturday. 

 The new agent aims to “take as little human attention as possible, and give as much value as possible,” DeepSeek said in a user manual on GitHub.

 Earlier this year, DeepSeek created the Harness team to focus on helping build AI agents that can take on coding services offered by leading US labs including Anthropic. 

 

 The term harness refers to the software scaffolding that wraps around a large language model which enables AI agents to manage more complex tasks. Anthropic and other developers have designed harnesses to make their software better at streamlining coding and other work.

 Chinese firms are locked in a heated rivalry with US developers to build more advanced — and affordable — AI tools. While OpenAI’s ChatGPT chatbot helped kick off the global AI race, many companies have since focused more on the lucrative emerging market for AI agents that can help automate work for developers and business professionals.   

 To bulk up its war chest and compete with Anthropic and OpenAI, DeepSeek has raised about 100 billion yuan ($15 billion) in its latest fundraising round, one of China’s largest-ever, ahead of its initial public offering.