DeepSeek unveils AI ‘super agent’ in coding race with OpenAI, Anthropic

The most striking detail is that DeepSeek's new 'Super Agent' claims autonomous goal adjustment, a capability not yet evident in rival systems. The announcement positioned DeepSeek alongside established players in the generative‑coding market. FourWeekMBA followed with a technical note that the agent can process multiple GitHub issues in parallel, suggesting a productivity boost for developers.

MarkTechPost added that DeepSeek released the dsh‑superagent code as an open‑source router‑worker model, inviting community contributions. Bloomberg highlighted the strategic intent, stating the enhanced agent is meant to rival Anthropic’s coding suite, while the South China Morning Post framed the launch as part of an escalating coding race that also includes OpenAI. The coverage diverges on the agent’s market positioning: Bloomberg portrays it as a direct competitor to Anthropic, whereas FourWeekMBA focuses solely on its parallel‑issue capability, leaving the competitive narrative ambiguous.

DeepSeek has now open‑sourced the super‑agent, enabling developers to experiment with the router‑worker framework. The current state is a publicly available codebase that signals DeepSeek’s entry into the AI‑driven coding arena, while industry observers await concrete performance benchmarks.