LLM WORLD

Qoder public preview

Alibaba Qwen · · Coding harness

Alibaba launched Qoder, an agentic coding platform, into free global public preview. It pairs a proprietary Next-Edit-Suggestion model with third-party frontier models — Claude, Gemini and GPT — behind a scheduler that picks one per task, and splits work between an Agent Mode for small edits and a Quest Mode that runs an end-to-end task autonomously.

Why it matters

The first Chinese lab to ship a coding agent aimed at developers outside China from day one, and the one that got the scale: reported at 100,000 users in five days and around five million by mid-2026. It is also the batch's clearest admission that the harness and the model are separable — Alibaba's own agentic IDE routes to Anthropic, Google and OpenAI models alongside its own.

This is LLM World's historical reading, not a claim by Alibaba Qwen. The facts above and the sources below are what it rests on.

Sources

Record status: verified · confidence 70/100

The launch press release, issued by Qoder and datelined Singapore, Thursday 21 August 2025, states in its own words that the public preview opened that day. Counted as primary because the release is the company's own statement, with the caveat that it was read on the ACCESS Newswire distribution site rather than on an Alibaba or qoder.com page — qoder.com/blog lists a 'Qoder Introduction' post but renders no publication dates, so the company's own site could not be used to confirm. SCOPE: this dates the free public preview, not general availability or the paid tiers. COMPANY: recorded as Alibaba Qwen to match this dataset's only Alibaba string; Qoder ships under its own brand and descends from the Alibaba Cloud Tongyi Lingma team, not the Qwen model team.