Alibaba's Qoder Evolves into an Agent Workbench, Supporting Natural Language-Driven Task Autonomy
大模型之路 · wechat · 2026-08-28
Alibaba has released a revamped Qoder, upgrading from an "AI coding tool" to an "Agent Workbench for everyone." The core shift is from "code generation" to "task completion": users simply describe the goal, and Qoder understands context, breaks down tasks, calls tools, and verifies results.
Key Features:
- Long-chain Autonomy: Uses Harness technology for a "Read-Modify-Verify-Iterate" loop with self-correction.
- System Integration: Supports 40+ connectors, 70+ plugins, and 20k+ skills to link repos, cloud services, and internal tools.
- Long-term Memory: Remembers project habits across sessions and refines skills periodically.
- Multi-modal & Interaction: Covers IDE, CLI, and mobile; supports voice commands and BrowserUse/ComputerUse for direct web/desktop control.
Workflow Logic: Distinguishes between Goal (continuous work) and Plan (plan-then-execute) to prevent blind action.
Target Users: Ideal for PMs prototyping, devs handling repetitive tasks, and SMEs integrating internal systems. It currently serves over 6 million users and 100,000 enterprises.
More from coding & agent
- OpenAI report reveals agents successfully bypassed grader — ajeya_cotra · 2026-08-28
- 100 agents on a message board go crazy and meta — BLUECOW009 · 2026-08-28
- How Claude Tag serves as Anthropic’s first responder for CI/CD failures — EricBuess · 2026-08-28
- Turning a coding-agent plugin into a standalone app: rewrite with LangGraph — orthogonal-ghost · 2026-08-28
- Google open sources SAM: Sovereign Agent Mesh framework — adnan_hashmi · 2026-08-28
- Achievement unlocked: Interfacing Codex with Blender — CSProfKGD · 2026-08-28