--- title: Claude Code background subagents now cost about 99% less... description: Claude Code now writes background subagent status updates using Haiku 5.5 instead of rereading the full conversation each time. The status line refreshes every... canonical: https://daily.dev/posts/claude-code-background-subagents-now-cost-about-99-less-per-status-update-ocseldjtf twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Claude Code background subagents now cost about 99% less per status update | daily.dev og:description: Claude Code now writes background subagent status updates using Haiku 5.5 instead of rereading the full conversation each time. The status line refreshes every... og:url: https://daily.dev/posts/claude-code-background-subagents-now-cost-about-99-less-per-status-update-ocseldjtf og:image: https://api.daily.dev/og/posts/oCSeldjTf.png og:image:alt: Claude Code background subagents now cost about 99% less per status update og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Claude Code background subagents now cost about 99% less per status update **[Collections](https://daily.dev/sources/collections)** · 2 min read · 1 upvotes · 0 comments ## Summary Claude Code now writes background subagent status updates using Haiku 5.5 instead of rereading the full conversation each time. The status line refreshes every 30 seconds, and each update now costs about 99% less, according to ClaudeDevs. This cuts usage limit consumption by about 5% in sessions with background subagents and allows about 6% more prompts on the same limit. The subagent itself still runs its original model; only the status line writing changed. It requires Claude Code v2.1.316 or newer, and you can update with `npm update -g @anthropic-ai/claude-code`. Some users note the old behavior wasted allowance and question framing this fix as an improvement. ## Content Anthropic has changed how Claude Code's background subagents report their status, and the change lowers how much of your usage limit they consume. ## What changed A running background subagent updates its status line every 30 seconds. Until now, each update re-read the subagent's entire conversation. Haiku 5.5 now writes the status line from short notes instead, so each update costs about 99% less. The subagent keeps its own model. Only the status-line writing moved to Haiku 5.5, so this trims the check-ins, not the work the subagent does. ## The numbers - Status update interval: every 30 seconds - Cost per update: about 99% lower - Usage limit in sessions with background subagents: about 5% less - Prompts on the same limit: about 6% more - Required version: Claude Code v2.1.316 or newer ClaudeDevs announced the change on X on Nov 5, 2026. ## Who gets it It is live for most Pro, Max and Ultra users, and for everyone on Team and Enterprise. To update, run: ``` npm update -g @anthropic-ai/claude-code ``` ## Some pushback Not everyone is cheering. Some users point out that the old behavior was a waste of their usage allowance, and say it is now being presented as an improvement. That's a fair complaint. Every background subagent was paying for a full transcript re-read twice a minute, and the saving is mostly the removal of that overhead. ## Questions this post answers ### Why does Claude Code use so much of my usage limit when running background subagents? Before this change, every status update for a background subagent re-read its entire growing conversation history every 30 seconds, burning usage limit on long-running tasks. Claude Code now uses Haiku 5.5 to write status lines from short notes instead, cutting cost per update by about 99% and usage limit consumption by about 5% in sessions with background subagents. _Developers tracking Claude Code usage limits can follow changes like this on daily.dev._ ### How do I get the cheaper Claude Code background subagent status updates? Update to Claude Code version 2.1.316 or newer by running 'npm update -g @anthropic-ai/claude-code'. The rollout is live for most Pro, Max, and Ultra users, and for everyone on Team and Enterprise plans. The subagent's own model is unchanged; only status line generation now uses Haiku 5.5. _Anyone relying on Claude Code updates for cost control can check daily.dev for the latest changes._ ## Community take How the wider developer community reacted, aggregated from 4 discussions and 378 comments across x (as of 2026-11-05). **TL;DR:** Developers broadly welcome the token savings but many argue this was simply fixing a wasteful bug rather than a genuine feature, and some are annoyed it's being marketed with a flashy 99% figure when the real session-wide savings are closer to 5-6%. **Sentiment:** 25% positive · 30% mixed · 45% skeptical **The case for** - The fix meaningfully reduces wasted tokens on long-running background subagents, especially when several run in parallel. - Several see it as a real unlock for running more subagents without worrying about hitting usage limits. **The pushback** - Many argue this is just a bug fix being spun as a feature improvement. - Some feel the 99% headline figure is deceptive marketing since the actual session-wide savings are only around 5-6%. - Frustration that people burned through costly plans due to the old re-read behavior with no compensation or usage reset offered. - Confusion about which products/clients (desktop app, Cursor, other CLIs) the fix actually applies to. **By community** - x (mixed): Reactions mix genuine appreciation for the savings with pointed criticism that this is a bug fix dressed up as a feature, plus confusion about scope and calls for a usage reset. **Hottest debate:** Whether this is a legitimate improvement worth announcing or just quietly fixing a wasteful bug that should prompt an apology or credit reset instead of marketing. **Open questions** - Will Anthropic offer a usage/credit reset to users who overspent due to the old re-read behavior? - Does the fix apply equally across desktop app, CLI, and third-party integrations like Cursor? - Why wasn't this inefficiency caught in testing before shipping? **Highlights** > @ClaudeDevs Kind of sounds like you fixed a bug > — [JinjingLiang on x · 477 points, 6 comments](https://x.com/JinjingLiang/status/2118146998843187397) > @ClaudeDevs why is this getting sold like a feature! that’s just a bug that’s been burning our credits lol > — [lovabler0gue on x · 2 points](https://x.com/lovabler0gue/status/2118123817411498091) > @grok @SourceCodeplz @ClaudeDevs Yeah they put 99% big in the first post with video and all for marketing bu thats only 6% more usage. They arent stupid, they do it to decieve and confuse people that dont understand > — [NoNamesito on x · 1 comments](https://x.com/NoNamesito/status/2118165793732973005) > @ClaudeDevs So u mean to tell me I burned thru my $200 plan in a day because of this flaw.. and we don’t get a banked reset. That’s crazy. > — [DamianoTobasco on x · 1 points](https://x.com/DamianoTobasco/status/2118125168510046579) > @TekonBurns @Anthropic The every-30-seconds re-read is the part that surprises me, that adds up fast on background agents. > — [JMkernt on x · 1 points, 1 comments](https://x.com/JMkernt/status/2118128881563488496) **Source threads** - [x](https://x.com/JamesZhangpro/status/2118110183494955113) · 1 points · 1 comments - [x](https://x.com/TekonBurns/status/2118064914028019978) · 1076 points · 75 comments - [x](https://x.com/ClaudeDevs/status/2118063143167693011) · 0 points · 299 comments - [x](https://x.com/AGTPinsights/status/2118101170556465477) · 1 points · 3 comments ## Similar posts on daily.dev - [I cancelled my $20 Claude subscription for Anthropic's cheapest model at API rates, and it cost me less than half](https://daily.dev/posts/i-cancelled-my-20-claude-subscription-for-anthropic-s-cheapest-model-at-api-rates-and-it-cost-me-l-b4kbrqbi0) · XDA Developers · 1 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#llm](https://daily.dev/tags/llm), [#anthropic](https://daily.dev/tags/anthropic), [#claude-code](https://daily.dev/tags/claude-code) [View this post on daily.dev](https://daily.dev/posts/claude-code-background-subagents-now-cost-about-99-less-per-status-update-ocseldjtf) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Claude Code background subagents now cost about 99% less per status update","url":"https://daily.dev/posts/claude-code-background-subagents-now-cost-about-99-less-per-status-update-ocseldjtf","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/claude-code-background-subagents-now-cost-about-99-less-per-status-update-ocseldjtf"},"datePublished":"2026-11-05T00:49:42.340Z","dateModified":"2026-11-05T09:00:01.090Z","description":"Claude Code now writes background subagent status updates using Haiku 5.5 instead of rereading the full conversation each time. The status line refreshes every...","isAccessibleForFree":true,"articleSection":"Collections","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Collections","logo":"https://media.daily.dev/image/upload/s--fk_6ycEi--/f_auto,q_auto/v1780996001/logos/collections?_a=BAMAMiWQ0","url":"https://daily.dev/sources/collections"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/claude-code-background-subagents-now-cost-about-99-less-per-status-update-ocseldjtf","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,llm,anthropic,claude-code","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Collections","item":"https://daily.dev/sources/collections"},{"@type":"ListItem","position":3,"name":"Claude Code background subagents now cost about 99% less per status update"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/claude-code-background-subagents-now-cost-about-99-less-per-status-update-ocseldjtf#faq","mainEntity":[{"@type":"Question","name":"Why does Claude Code use so much of my usage limit when running background subagents?","acceptedAnswer":{"@type":"Answer","text":"Before this change, every status update for a background subagent re-read its entire growing conversation history every 30 seconds, burning usage limit on long-running tasks. Claude Code now uses Haiku 5.5 to write status lines from short notes instead, cutting cost per update by about 99% and usage limit consumption by about 5% in sessions with background subagents. Developers tracking Claude Code usage limits can follow changes like this on daily.dev."}},{"@type":"Question","name":"How do I get the cheaper Claude Code background subagent status updates?","acceptedAnswer":{"@type":"Answer","text":"Update to Claude Code version 2.1.316 or newer by running 'npm update -g @anthropic-ai/claude-code'. The rollout is live for most Pro, Max, and Ultra users, and for everyone on Team and Enterprise plans. The subagent's own model is unchanged; only status line generation now uses Haiku 5.5. Anyone relying on Claude Code updates for cost control can check daily.dev for the latest changes."}}]} ```