DevConsensus
Models / Claude Sonnet 4.6
Model · HN opinions Nov 2025 – Nov 2026

Claude Sonnet 4.6: what developers actually say

We read every Hacker News comment from Nov 2025 – Nov 2026 that gave a real opinion on Claude Sonnet 4.6. 59 of the 65 opinions come from people who say they use it themselves.

65%approval, all opinions
–approval, last 90 days
#30of 45 models ranked
66different developers

Developers give Claude Sonnet 4.6 a 65% approval score across 65 opinions from 66 authors, and 59 of those opinions come from firsthand use. Code quality draws the most attention at 20 positive and 13 negative mentions, while speed is the clearest strength at 7 positive and 1 negative. Praise focuses on everyday coding, catching errors on a second pass, and full-stack work. Discussion volume fell from 20 opinions in April 2026 to 3 in July and 1 in each of August and September, so the recent trend rests on thin data, and July approval was 17%.

Pick it if

  • Want fast help with everyday coding
  • Need a model that catches its own errors on a second pass
  • Run full-stack builds or prototypes with light supervision

Look elsewhere if

  • Point agents at production data without guardrails
  • Need a model that reviews its own code fairly
  • Face tight weekly usage caps

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
13
20
Speed
1
7
Reliability
3
3
Usage limits
2
3
Autonomy
1
3
Cost & pricing
1
2

How opinion has moved

050100Feb '26Mar '26Apr '26May '26Jun '26Jul '26
Monthly approval (line) and number of opinions (bars). Months with fewer than 3 opinions are left out.

What developers say

Praise

Says Sonnet 4.6 performs better than Opus 5 in their experience.HN comment · Aug 16, 2026
Says Sonnet 4.6 caught and corrected all the errors in a second pass.HN comment · Jun 30, 2026
Says Sonnet 4.6 handled basic full-stack coding without trouble.HN comment · Jun 13, 2026
Says Sonnet 4.6 is better than Opus 4.8.HN comment · Jun 8, 2026
Says Sonnet 4.6 sped up a spec-to-code project so step 8 of 9 took about 12 hours.HN comment · Jun 6, 2026
Says Claude Sonnet 4.5 and 4.6 did most of the heavy lifting in building their Android app.HN comment · May 31, 2026
Says Sonnet 4.6 is good enough for vibe coding, so they stick with it over newer models.HN comment · May 28, 2026
Switched back to Sonnet 4.6 in Claude Code after repeatedly regretting trying Opus 4.7.HN comment · May 23, 2026

Complaints

Says Claude Sonnet 4.6 has recently felt weak compared with newer models.HN comment · Jul 9, 2026
Says Kimi K2.6 and DeepSeek V4 Pro beat Sonnet 4.6, which still trails Opus.HN comment · Jul 1, 2026
Says Sonnet 4.6 is clearly weaker than MiMo V2.5 Pro for their serious work.HN comment · Jun 10, 2026
Ran into Sonnet 4.6 via an old Claude Code plugin and found it surprisingly dumb, highlighting recent progress.HN comment · May 28, 2026
Says Sonnet 4.6 would have exhausted their weekly Max limits for weeks, which is why they switched.HN comment · May 15, 2026
Uses Sonnet 4.6 regularly but must watch it closely, since agents sometimes delete tests or nuke databases.HN comment · May 9, 2026
Says Sonnet 4.6 rates its own code unfairly when the same model both writes and judges it.HN comment · May 3, 2026
Says Sonnet 4.6 is significantly worse than a year ago, and their employer requires it exclusively.HN comment · Apr 24, 2026

Summaries are paraphrased. Each links to the original comment.

Claude Sonnet 4.6 compared

FAQ

Is Claude Sonnet 4.6 worth it?

It scores 65% approval across 65 opinions, so most developers find it useful. Some call it a great everyday agent that is barely better than cheaper models. Two people preferred GLM 5.1 instead and two preferred DeepSeek V4 Pro.

Is Claude Sonnet 4.6 better than Opus?

One person preferred Sonnet 4.6 over Opus 4.7 and switched back after regretting the upgrade. Two preferred it over Opus 4.6, one over Opus 5 and one over Opus 4.8. The data is too thin to call a clear winner.

Does Claude Sonnet 4.6 hit usage limits?

Usage limits drew 3 positive and 2 negative mentions. One developer said the model would have exhausted their weekly Max limits for weeks, which led them to switch. Another said Claude's window runs out after about 45 minutes.