Cursor: what developers actually say
We read every Hacker News comment from Nov 2025 – Nov 2026 that gave a real opinion on Cursor. 583 of the 766 opinions come from people who say they use it themselves.
Developers give Cursor a 55% approval score across 766 opinions from 600 authors, and 53% over the last 90 days. They like having Claude, Codex and other models in one interface, plus fast autocomplete and a useful review UI. Reliability is the weak point, with 14 positive mentions against 58 negative, while cost is split at 58 positive and 60 negative. The recent trend is roughly flat, with monthly approval swinging from 70% in September to 44% in October on small samples.
Pick it if
- Want Claude, Codex and other agents in one interface
- Care most about autocomplete speed
- Want to switch models inside the editor
Look elsewhere if
- Need dependable agent behavior
- Have a tight budget or hit usage caps
- Depend on extensions Cursor does not support
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.
How opinion has moved
What developers say
Praise
Complaints
Summaries are paraphrased. Each links to the original comment.
Cursor compared
Best Cursor alternatives, according to developers who switched →
FAQ
Is Cursor worth it?
Approval is 55% overall and 53% over the last 90 days. Users praise the unified interface, speed and autocomplete, but reliability and pricing draw frequent complaints. It is a reasonable pick if those tradeoffs fit your workflow.
Is Cursor better than Claude Code?
In this sample, 61 developers preferred Claude Code instead of Cursor, while 42 preferred Cursor over Claude Code. That favors Claude Code on head-to-head preference, though the gap is not overwhelming.
Why do developers stop using Cursor?
The largest negative gap is reliability, at 14 positive mentions and 58 negative. Cost is evenly split at 58 and 60, and some users cite missing extensions, a clunky feel, or concerns about the model lineup.