Claude Opus 4.6: loved for code quality, held back by reliability.
287 Hacker News opinions from 257 developers, Nov 2025 – Nov 2026. 242 come from people who say they use it. Here is what they praise, what they complain about, and who switched.
Where Claude Opus 4.6 is strong — and where it is not.
Claude Opus 4.6 earns 63% approval across 287 opinions. Developers praise its code quality, while reliability is the recurring complaint. In the last 90 days it sits at 63%.
Pick it if you
- Want strong code quality, the top aspect at 78 positive to 43 negative
- Value autonomy, which scored 21 positive to 8 negative
- Have well-specified tasks it can finish in one pass
Look elsewhere if you
- Need dependable output, since reliability ran 19 positive to 39 negative
- Work in niche languages or platform APIs that users report it struggles with
- Need predictable speed or cost, both of which scored near neutral
Where it wins and loses.
Mentions across 287 opinions. Bars are counts, not votes; a tool can be both praised and criticized on the same topic.
View data table
| Topic | Praise | Complaints | Mentions |
|---|---|---|---|
| Code quality | 78 | 43 | 121 |
| Autonomy | 21 | 8 | 29 |
| Cost & pricing | 7 | 7 | 14 |
| Speed | 4 | 6 | 10 |
| Reliability | 19 | 39 | 58 |
How opinion has moved.
Monthly approval, Feb 2026 – Sep 2026. Bars show opinion volume; low-volume months are still included but faded.
View data table
| Month | Approval | Opinions |
|---|---|---|
| Feb 2026 | 73% | 66 |
| Mar 2026 | 58% | 76 |
| Apr 2026 | 55% | 74 |
| May 2026 | 61% | 23 |
| Jun 2026 | 82% | 14 |
| Jul 2026 | 82% | 11 |
| Aug 2026 | 54% | 14 |
| Sep 2026 | 71% | 7 |
What developers say.
Summaries are paraphrased. Each links to the original Hacker News comment.
Praise
4 selectedOpus 4.6 found more issues in their code than they did, catching semantic bugs linters miss.
Opus 4.6 noticeably raised accuracy for FreeCAD and OpenSCAD code, often one-shotting described geometry.
Opus 4.6 is best at getting to the point and doing what is asked.
Their system prompt worked perfectly with Opus 4.6.
Complaints
4 selectedOpus 4.6 struggled with standard GHC Haskell, needing more turns than for Java or TypeScript.
They tried Opus 4.6 for coding but gave up on it.
Opus 4.6 got stuck in local minima and ignored instructions during a costly $600 experiment.
Opus 4.6 struggled with macOS C APIs, such as CoreGraphics coordinate inversion, when they last tried it.
Questions developers ask.
Is Claude Opus 4.6 worth it?
Is Claude Opus 4.6 better than GPT-5.4?
Is Claude Opus 4.6 good for large codebases?
The data behind this page.
Developers give Claude Opus 4.6 a 63% approval score across 287 opinions from 257 authors, and 242 of those opinions are firsthand. Code quality is the main draw at 78 positive mentions against 43 negative, and autonomy follows at 21 to 8. Reliability is the clear weak point at 19 positive against 39 negative, with users describing degraded behavior, dropped tests and long stalls. Monthly approval ran 73 in February, 55 in April, 82 in June and July, 54 in August, 71 in September and 100 from only 2 opinions in October, and the 90 day figure is also 63.
Every number comes from public Hacker News comments collected through the HN Search API. A language model tagged each comment for sentiment, first-hand use, topic and switches; a human spot-checks samples. No vendor can pay for placement.
Get the monthly report
Which AI coding tools gained or lost developer approval on Hacker News, and why people switched. One email a month.