Terminal agent · Alibaba · HN opinions Nov 2025 – Nov 2026

Qwen Code: what developers actually say

We read every Hacker News comment from Nov 2025 – Nov 2026 that gave a real opinion on Qwen Code. 17 of the 19 opinions come from people who say they use it themselves.

63%approval, all opinions

–approval, last 90 days

#16of 22 tools ranked

16different developers

Qwen Code holds 63% approval across 19 opinions from 16 authors, and 17 of those opinions come from firsthand use. Cost is the clearest strength with 4 positive cost comments and no negative ones, including a report of about 300M free tokens a day. Reliability is the weak spot at 1 positive and 4 negative comments, with reports of failed write_file calls, lost work after token exhaustion, and subagents overloading a shared local server. Monthly approval fell to 0% in July and October 2026, though each of those months rests on a single opinion.

Pick it if

Want a generous free daily token quota
Run models locally and want no queues or token dependence
Care more about cost per result than top-tier model quality

Look elsewhere if

Need dependable file writes and tool calls
Share a local LLM server with other users
Need the strongest model available for hard tasks

Visit Qwen Code ↗

Strengths and weaknesses

Each row counts opinions that mentioned the topic. Bars to the right of the line are praise, bars to the left are complaints.

PraiseComplaints

Reliability

Cost & pricing

Usage limits

Code quality

How opinion has moved

Not enough monthly data for a trend yet.

What developers say

Praise

Says Qwen Code feels faster and nicer to work with than Claude, though less smart.HN comment · May 17, 2026

Used Qwen CLI on a hobby project; it finished 95%, slightly below top Anthropic and OpenAI models.HN comment · Apr 16, 2026

Says Qwen Code's free daily quota is very generous, around 300M tokens a day for them.HN comment · Apr 3, 2026

Reports good cost-benefit results using the Qwen CLI with its coder model in the cloud.HN comment · Mar 8, 2026

Says Qwen Code CLI works well with a Qwen model, after being unhappy with other local tools.HN comment · Feb 21, 2026

Uses Qwen Code CLI daily for most dev work, valuing local use with no queues or token dependence.HN comment · Feb 20, 2026

Says Qwen Code paired with Qwen3-Coder-Next is pretty capable.HN comment · Feb 5, 2026

Says Qwen Code showed better reasoning and structure than Gemini CLI in their trial.HN comment · Nov 18, 2025

Complaints

Says Qwen Code immediately launched about six parallel subagents, overloading a shared local LLM server.HN comment · Oct 9, 2026

Says Qwen Code still failed on write_file tool calls when using Qwen3.6 27B.HN comment · Jul 12, 2026

Lost work with Qwen Code when it got confused after running out of tokens.HN comment · May 12, 2026

Says Codex is about twice as good as Qwen Code.HN comment · Mar 10, 2026

Says Qwen tends to swing back and forth too much.HN comment · Feb 26, 2026

Qwen Coder CLI was among the tools that hit per-minute request limits almost immediately.HN comment · Nov 8, 2025

Summaries are paraphrased. Each links to the original comment.

FAQ

Is Qwen Code worth it?

It depends on your priorities. Approval is 63% from 19 opinions. Developers praise the cost and the free quota, while reliability drew 4 negative comments against 1 positive.

Is Qwen Code better than Claude Code?

One developer preferred Claude Code. Another said Qwen Code feels faster and nicer to work with than Claude but is less smart. One comment put Codex at about twice as good as Qwen Code.

Does Qwen Code have a free tier?

One developer reported a free daily quota of around 300M tokens. Usage limits drew 2 positive comments and 1 negative, and one developer said the CLI hit per-minute request limits almost immediately.