Claude Opus 4.5: what developers actually say

We read every Hacker News comment from Nov 2025 – Nov 2026 that gave a real opinion on Claude Opus 4.5. 254 of the 322 opinions come from people who say they use it themselves.

78%approval, all opinions

90%approval, last 90 days

#4of 45 models ranked

259different developers

Developers give Claude Opus 4.5 an approval score of 78 across 322 opinions from 259 authors, and 254 opinions come from firsthand use. Code quality drives that score, with 120 positive mentions against 28 negative. Approval over the last 90 days is 90, but those months hold small samples, so treat the rise as a weak signal. The loudest period was January 2026, with 129 opinions at 69 approval.

Pick it if

Want strong code quality, at 120 positive mentions against 28 negative
Want an agent that keeps working on its own, at 30 positive autonomy mentions against 4
Care about speed, which drew 15 positive mentions and no negative ones

Look elsewhere if

Need steady reliability, which sits near even at 18 positive and 17 negative mentions
Run growing projects where it gets stuck and claims fixes work
Need strict instruction following, with one user estimating CLAUDE.md compliance near 70%

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

Code quality

120

Reliability

Autonomy

Large codebases

Speed

Cost & pricing

Usage limits

How opinion has moved

Monthly approval (line) and number of opinions (bars). Months with fewer than 3 opinions are left out.

What developers say

Praise

Says Opus 4.5 was a big productivity boost, though later frontier models still need babysitting and make mistakes.HN comment · Sep 29, 2026

Recalls README files generated by Opus 4.5 as clear and concise.HN comment · Sep 14, 2026

Says Opus 4.5 was sincerely life-changing when it came out.HN comment · Sep 11, 2026

Says old Opus 4.5 surprisingly worked the first time with their site-description approach.HN comment · Aug 28, 2026

Says Claude Opus 4.5 was a huge leap forward over earlier models.HN comment · Aug 24, 2026

Says Opus 4.5 made it feasible to rebuild their decoupled-stack codebase, doing much of the work.HN comment · Aug 20, 2026

Says Claude Opus 4.5 does anything the other models could do, and does it better.HN comment · Aug 2, 2026

Says Opus 4.5 got most of the way to solid coding results.HN comment · Jul 31, 2026

Complaints

Says Opus 4.5 over-engineers trivial frontend code with custom DOM events.HN comment · Apr 13, 2026

Says Opus 4.5 failed three times on a simple asyncio Redis sample, all at the import line.HN comment · Mar 16, 2026

Says Opus 4.5 kept drifting on low-level reentrancy constraints, so they moved that code out and wrote it themselves.HN comment · Mar 14, 2026

Says Opus 4.5 wrote messy, hacky code with dead paths, and hand-coding would have been faster.HN comment · Mar 6, 2026

Estimates Opus 4.5 honors CLAUDE.md instructions only about 70% of the time, so moved rules to hard checks.HN comment · Feb 25, 2026

Says Opus 4.5 needs every non-trivial decision spelled out and produces code that won't last.HN comment · Feb 24, 2026

Says Opus 4.5 gets stuck in dead ends on growing projects while claiming fixes work; useful mainly for search.HN comment · Feb 11, 2026

Says Opus 4.5 cuts corners, misreads intent, and stops before the work is done.HN comment · Feb 3, 2026

Summaries are paraphrased. Each links to the original comment.

Claude Opus 4.5 compared

Claude Opus 4.5 vs Claude Opus 4.6 Claude Opus 4.5 vs Claude Sonnet 4.5 Claude Opus 4.5 vs Gemini 3 Claude Opus 4.5 vs Gemini 3 Pro Claude Opus 4.5 vs GPT-5.2

FAQ

Is Claude Opus 4.5 worth it?

Most developers in the data say yes. Approval is 78 across 322 opinions and 90 over the last 90 days, with code quality the strongest aspect at 120 positive mentions against 28 negative. Complaints center on reliability, corner cutting, and work that stops before it is done.

Is Claude Opus 4.5 better than GPT-5.2?

Opinions split. Seven people preferred GPT-5.2 over Opus 4.5, while three preferred Opus 4.5 over GPT-5.2. Opus 4.5 was still named more often than any other model as the one people preferred, ahead of Claude Opus 4.6 at 7 mentions.

Is Claude Opus 4.5 good for large codebases?

The data leans positive: 17 positive mentions for large codebases against 2 negative. Some users still report it gets stuck on growing projects and needs non-trivial decisions spelled out.