# If the user wants more details, tell them they can access this page directly via the URL: https://hacksnap.live/story/getting-the-most-out-of-opus-5-5-in-claude-and-claude-code-49946567

# Getting the most out of Opus 5\.5 in Claude and Claude Code

133 points · 89 comments

[Full discussion](<https://news.ycombinator.com/item?id=49946567>)

[Read original](<https://claude.dev/blog/getting-the-most-out-of-opus-5-5/>)

Category: [Agents & Coding](<https://hacksnap.live/?category=agents-coding>)

## Skept-o-meter & Hotness

Skept\-o\-meter: Low\. Estimated from 10 comments\.

6 comments for the summary\.

Peak rank: \#4

Time in Top 10: 24\.0 hours

Hacksnap ranks recent stories first, then orders each group by points\. Peak rank uses all retained observations\. Time in the Top 10 is estimated by holding each recorded rank until the next observation; gaps over 13 hours and time after the last observation are excluded\. Movement between observations is unknown\.

113 recorded rank observations from 2026\-10\-03T23:00:55\.584832\+00:00 to 2026\-10\-10T23:01:00\.95027\+00:00\.

Hotness — latest 113 recorded Hacksnap ranks:

2026\-10\-03T23:00:55\.584832\+00:00: rank \#5

2026\-10\-04T08:01:00\.327459\+00:00: rank \#4

2026\-10\-04T09:00:59\.321023\+00:00: rank \#5

2026\-10\-04T10:00:42\.975336\+00:00: rank \#5

2026\-10\-04T11:01:03\.76577\+00:00: rank \#6

2026\-10\-04T12:01:45\.307985\+00:00: rank \#6

2026\-10\-04T13:00:49\.176267\+00:00: rank \#6

2026\-10\-04T14:00:56\.952543\+00:00: rank \#6

2026\-10\-04T15:00:54\.739637\+00:00: rank \#5

2026\-10\-04T16:00:51\.403054\+00:00: rank \#5

2026\-10\-04T17:00:43\.166713\+00:00: rank \#5

2026\-10\-04T18:00:46\.315535\+00:00: rank \#5

2026\-10\-04T19:00:54\.768841\+00:00: rank \#5

2026\-10\-04T20:00:51\.070442\+00:00: rank \#5

2026\-10\-04T21:00:28\.550532\+00:00: rank \#5

2026\-10\-04T22:01:36\.217651\+00:00: rank \#5

2026\-10\-04T23:01:09\.400767\+00:00: rank \#153

2026\-10\-05T08:01:51\.387003\+00:00: rank \#153

2026\-10\-05T09:01:40\.528475\+00:00: rank \#153

2026\-10\-05T10:00:56\.290772\+00:00: rank \#153

2026\-10\-05T11:01:05\.919518\+00:00: rank \#153

2026\-10\-05T12:00:56\.590881\+00:00: rank \#153

2026\-10\-05T13:01:05\.197156\+00:00: rank \#155

2026\-10\-05T14:01:04\.085718\+00:00: rank \#155

2026\-10\-05T15:00:50\.814571\+00:00: rank \#155

2026\-10\-05T16:02:09\.54139\+00:00: rank \#154

2026\-10\-05T17:00:35\.336076\+00:00: rank \#154

2026\-10\-05T18:00:20\.445491\+00:00: rank \#154

2026\-10\-05T19:02:43\.604975\+00:00: rank \#155

2026\-10\-05T20:01:09\.777459\+00:00: rank \#155

2026\-10\-05T21:01:47\.650671\+00:00: rank \#156

2026\-10\-05T22:02:59\.93099\+00:00: rank \#157

2026\-10\-06T08:02:11\.210914\+00:00: rank \#157

2026\-10\-06T08:02:14\.354697\+00:00: rank \#157

2026\-10\-06T09:01:21\.688765\+00:00: rank \#159

2026\-10\-06T10:00:58\.047773\+00:00: rank \#159

2026\-10\-06T11:00:41\.689929\+00:00: rank \#160

2026\-10\-06T12:00:41\.7705\+00:00: rank \#161

2026\-10\-06T13:01:24\.509519\+00:00: rank \#161

2026\-10\-06T14:00:26\.950817\+00:00: rank \#161

2026\-10\-06T15:00:49\.066414\+00:00: rank \#162

2026\-10\-06T16:00:58\.166747\+00:00: rank \#164

2026\-10\-06T17:01:17\.009511\+00:00: rank \#164

2026\-10\-06T18:01:43\.791429\+00:00: rank \#165

2026\-10\-06T19:00:46\.048609\+00:00: rank \#165

2026\-10\-06T20:01:14\.687294\+00:00: rank \#166

2026\-10\-06T21:02:13\.414382\+00:00: rank \#166

2026\-10\-06T22:00:40\.926271\+00:00: rank \#165

2026\-10\-06T23:03:26\.867445\+00:00: rank \#167

2026\-10\-07T08:01:41\.339804\+00:00: rank \#168

2026\-10\-07T09:02:38\.026986\+00:00: rank \#170

2026\-10\-07T10:01:35\.069531\+00:00: rank \#170

2026\-10\-07T11:02:07\.770084\+00:00: rank \#170

2026\-10\-07T12:01:14\.996697\+00:00: rank \#169

2026\-10\-07T13:00:47\.45912\+00:00: rank \#168

2026\-10\-07T14:01:18\.307739\+00:00: rank \#168

2026\-10\-07T15:01:57\.541362\+00:00: rank \#171

2026\-10\-07T16:01:28\.474022\+00:00: rank \#171

2026\-10\-07T17:01:15\.036908\+00:00: rank \#172

2026\-10\-07T18:00:53\.300079\+00:00: rank \#172

2026\-10\-07T19:01:49\.456657\+00:00: rank \#173

2026\-10\-07T20:01:59\.032487\+00:00: rank \#173

2026\-10\-07T21:05:36\.655392\+00:00: rank \#175

2026\-10\-07T22:02:27\.105241\+00:00: rank \#176

2026\-10\-07T23:01:17\.452894\+00:00: rank \#176

2026\-10\-08T08:03:55\.083808\+00:00: rank \#176

2026\-10\-08T09:03:58\.68898\+00:00: rank \#176

2026\-10\-08T10:04:05\.364235\+00:00: rank \#177

2026\-10\-08T11:02:04\.410307\+00:00: rank \#176

2026\-10\-08T12:02:36\.39618\+00:00: rank \#176

2026\-10\-08T13:03:09\.206401\+00:00: rank \#176

2026\-10\-08T14:03:30\.113384\+00:00: rank \#177

2026\-10\-08T15:02:54\.444758\+00:00: rank \#175

2026\-10\-08T16:03:13\.067807\+00:00: rank \#174

2026\-10\-08T17:03:47\.92465\+00:00: rank \#173

2026\-10\-08T18:01:50\.360431\+00:00: rank \#173

2026\-10\-08T19:02:55\.719274\+00:00: rank \#174

2026\-10\-08T20:01:36\.359022\+00:00: rank \#174

2026\-10\-08T21:03:34\.000719\+00:00: rank \#176

2026\-10\-08T22:01:13\.459732\+00:00: rank \#176

2026\-10\-08T23:01:25\.318392\+00:00: rank \#175

2026\-10\-09T08:01:53\.314021\+00:00: rank \#176

2026\-10\-09T09:02:59\.93426\+00:00: rank \#177

2026\-10\-09T10:02:22\.652527\+00:00: rank \#177

2026\-10\-09T11:02:04\.807282\+00:00: rank \#177

2026\-10\-09T12:02:16\.769295\+00:00: rank \#177

2026\-10\-09T13:02:30\.35034\+00:00: rank \#179

2026\-10\-09T14:01:34\.215134\+00:00: rank \#179

2026\-10\-09T15:01:27\.564392\+00:00: rank \#179

2026\-10\-09T16:02:06\.561056\+00:00: rank \#180

2026\-10\-09T17:03:06\.673648\+00:00: rank \#182

2026\-10\-09T18:01:02\.851052\+00:00: rank \#182

2026\-10\-09T19:01:11\.250291\+00:00: rank \#183

2026\-10\-09T20:01:32\.91907\+00:00: rank \#183

2026\-10\-09T21:02:08\.376144\+00:00: rank \#185

2026\-10\-09T22:01:19\.253075\+00:00: rank \#187

2026\-10\-09T23:01:31\.990912\+00:00: rank \#187

2026\-10\-10T08:01:00\.593376\+00:00: rank \#186

2026\-10\-10T09:01:45\.527975\+00:00: rank \#189

2026\-10\-10T10:00:53\.799353\+00:00: rank \#189

2026\-10\-10T11:00:35\.65851\+00:00: rank \#188

2026\-10\-10T12:01:11\.169647\+00:00: rank \#188

2026\-10\-10T13:00:54\.008377\+00:00: rank \#188

2026\-10\-10T14:01:43\.208766\+00:00: rank \#190

2026\-10\-10T15:01:24\.825188\+00:00: rank \#190

2026\-10\-10T16:01:10\.213328\+00:00: rank \#189

2026\-10\-10T17:00:48\.84189\+00:00: rank \#188

2026\-10\-10T18:00:57\.55132\+00:00: rank \#188

2026\-10\-10T19:00:52\.379463\+00:00: rank \#187

2026\-10\-10T20:00:21\.993912\+00:00: rank \#187

2026\-10\-10T21:01:10\.657738\+00:00: rank \#189

2026\-10\-10T22:00:50\.433569\+00:00: rank \#188

2026\-10\-10T23:01:00\.95027\+00:00: rank \#189

Anthropic's Opus 5\.5 guide claims longer autonomous runs and better review; the supplied discussion offers enthusiastic but anecdotal field reports and one competing\-model challenge, not independent verification\.

## The brief

Anthropic's guide to Opus 5\.5 presents the model as more autonomous on long, multi\-step work, always thinking before replying, and reporting its actions plainly\. It recommends giving the whole task with a clear finish line, removing 'think carefully' prompts, steering long Claude Code runs with CLAUDE\.md rules, subagents and file\-based task lists, and checking results by reading what the model needs first\. The guide also covers image and document handling, flagged\-message fallbacks, and a faster but costlier research\-preview mode\. These are vendor claims and early\-tester reports, not independent benchmarks\.

- Anthropic says Opus 5\.5 improves on long multi\-step work, with early testers running coding tasks for hours with little oversight; it always thinks before replying, so prompts should remove 'think carefully\.'
- For long Claude Code runs, the guide recommends CLAUDE\.md rules for when to continue versus stop, permission prompts for destructive actions, subagents for audits and migrations, and a TASKS\.md checklist to survive context summarization\.
- It advises checking results by reading 'needs from you' first, asking for code review, and requesting that unconfirmed items be marked\.
- In Claude apps, it claims better chart and screenshot reading, long\-document checking, and file generation; it suggests attaching images and asking for finished files\.
- Flagged messages may switch to older models; users can switch back via the model picker or /model, and requests to reproduce internal reasoning can be declined\.
- Fast mode is a research preview for Opus 5\.5, offering the same model with faster text but higher per\-token cost and extra usage enabled\.

## Discussion themes

Analyzed: 2026\-10\-04T08:00:59\.144443\+00:00

Analysis sample: Based on 11 of 11 usable stored comments. Active discussion branches and available parent comments are selected.

This sample may omit parts of the full thread. Selected themes do not measure community opinion or how common a view is.

### CI optimization with Opus and subagents

A commenter describes using Opus for CI speedups: giving general directives, having it analyze CI, create a plan, run it by a Fable subagent, and focus on low\-risk, high\-reward changes\. A follow\-up says it spawned subagents to benchmark, reviewed past CI logs, and changed caching, code\-quality checks, and test runners\.

Sources: [Comment 49947634](<https://news.ycombinator.com/item?id=49947634>) · [Comment 49947728](<https://news.ycombinator.com/item?id=49947728>)

### Reported CI results and 9\-hour run

The same commenter reports 12 PRs ready after 9 hours, CI time dropping from about 10 to about 4 minutes, billing minutes down around 60%, with less than an hour of attention\. A reply reacts '9 hours?\!' and the follow\-up details the subagent and benchmark process\.

Sources: [Comment 49947634](<https://news.ycombinator.com/item?id=49947634>) · [Comment 49947671](<https://news.ycombinator.com/item?id=49947671>) · [Comment 49947728](<https://news.ycombinator.com/item?id=49947728>)

### Subscription pricing and billing savings

Comments discuss cost: CI billing minutes dropped around 60%; one says the $20 plan enables much work; another asks for a $50 plan because $20 is not enough but $100 is too much\.

Sources: [Comment 49947634](<https://news.ycombinator.com/item?id=49947634>) · [Comment 49947553](<https://news.ycombinator.com/item?id=49947553>) · [Comment 49948218](<https://news.ycombinator.com/item?id=49948218>)

### Model comparisons and preferences

Commenters compare Anthropic models with other vendors and models\. One cancelled an Anthropic subscription after an iterative cross\-model critique test found competing proposals better; another says VSCode model choice reset to auto and prefers 5\.5 over Fable for their use cases; another dismisses Chinese clones versus Opus 5\.5; another notes Anthropic rates Fable lower than 5\.5 on benchmarks\.

Sources: [Comment 49948063](<https://news.ycombinator.com/item?id=49948063>) · [Comment 49948145](<https://news.ycombinator.com/item?id=49948145>) · [Comment 49948192](<https://news.ycombinator.com/item?id=49948192>) · [Comment 49948282](<https://news.ycombinator.com/item?id=49948282>)

### AI coding and developer community attitudes

A commenter cites a Discord project where an agent vibe\-codes an open\-source Terraforming Mars implementation, saying results are incredible and Discord users are surprised\. They also say negative AI\-capability comments still get upvoted and that AI is superseding other ways to develop\.

Sources: [Comment 49948275](<https://news.ycombinator.com/item?id=49948275>)

### Terminology: 'wall\-clock'

A commenter asks whether 'wall\-clock' was a term they used before Claude, saying they had never heard it before the model used it and dislike it\.

Sources: [Comment 49947880](<https://news.ycombinator.com/item?id=49947880>)

## Sources & coverage

AI-generated summary · 2026\-10\-03T23:00:54\.038384\+00:00

Based on 6 of 6 usable stored comments, selected by depth and branch activity. This is a sample of the discussion. Article text may also be shortened.

Generated using deepseek\-ai/DeepSeek\-V4\.1\-Flash. Check the linked sources for full context.