DeepSeek unveils AI ‘super agent’ in coding race with OpenAI, Anthropic
DeepSeek has released dsh-superagent, an open-source coding tool designed to compete with Anthropic's Claude Code and OpenAI Codex. The tool utilizes a Router-Worker architecture that allows it to handle GitHub issues in parallel and adjust goals autonomously. Unlike traditional agents, the Super-Agent proactively identifies coding problems and suggests solutions without waiting for developer prompts. Performance data indicates the agent completes development tasks 1.9x faster at p95 compared to single-session agents, targeting enterprise teams and high-speed software development workflows.
Listen to Live Briefing
Real-time synthesized voice briefing · Live Feeds Desk
- ✓ DeepSeek released dsh-superagent as an open-source coding agent.
- ✓ The tool uses a Router-Worker architecture.
- ✓ The agent is designed to compete with Anthropic's Claude Code and OpenAI Codex.
- ✓ The Super-Agent can handle GitHub issues in parallel.
What changed
DeepSeek open-sourced the dsh-superagent, a proactive Router-Worker coding agent.
Live updates
-
DeepSeek launches dsh-superagent to rival Anthropic and OpenAI in AI coding
DeepSeek has released dsh-superagent, an open-source coding tool designed to compete with Anthropic's Claude Code and OpenAI Codex. The tool utilizes a Router-Worker architecture that allows it to handle GitHub issues in parallel and adjust goals autonomously. Unlike traditional agents, the Super-Agent proactively identifies coding problems and suggests solutions without waiting for developer prompts. Performance data indicates the agent completes development tasks 1.9x faster at p95 compared to single-session agents, targeting enterprise teams and high-speed software development workflows.
Why it matters
The release intensifies the race for autonomous AI agents capable of independent software engineering. By open-sourcing the architecture, DeepSeek challenges the proprietary models of leading US AI firms. This move shifts the focus from simple code generation to proactive project management.
What is confirmed
- DeepSeek released dsh-superagent as an open-source coding agent.
- The tool uses a Router-Worker architecture.
- The agent is designed to compete with Anthropic's Claude Code and OpenAI Codex.
- The Super-Agent can handle GitHub issues in parallel.
- The architecture allows for autonomous goal adjustment.
Still unconfirmed
- The agent completes dev tasks 1.9x faster at p95 than single-session agents.
- The tool suggests solutions without waiting for developer prompts.
What to watch next
- Independent benchmarks comparing dsh-superagent performance against Claude Code
- Adoption rates of the open-source architecture by enterprise development teams
confidence 90%Sources used for this update (12)
- www.scmp.com — The EU’s China dilemma: how to counter Beijing without alienating friends | South China Morning Post
- www.siliconrepublic.com — Silicon Republic | Technology, science and start-up news
- www.therundown.ai — The AI bubble meets Anthropic’s real numbers
- South China Morning Post — DeepSeek unveils AI ‘super agent’ in coding race with OpenAI, Anthropic
- Bloomberg.com — DeepSeek Drops Enhanced Coding Agent With Aim to Rival Anthropic
- MarkTechPost — DeepSeek Open-Sources dsh-superagent, a Router-Worker Coding Agent
- FourWeekMBA — DeepSeek Releases Agent That Works GitHub Issues in Parallel
- 헤럴드경제 — DeepSeek unveils 'Super Agent' architecture that handles goal adjustment autonomously
- www.straitstimes.com — DeepSeek launches Super-agent coding tool to rival Anthropic ...
- www.digitaltrends.com — DeepSeek wants to steal Claude Code and Codex’s lunch with an ...
- emergent.sh — DeepSeek Coding Agent Launches to Challenge Anthropic
- www.eneralabs.com — DeepSeek Super-Agent: Autonomous Coding for Enterprise Teams
Community Sentiment: How do you assess this situation?
Voice your perspective · Real-time aggregated sentiment from the Live Feeds community