--- title: Anthropic releases Opus 5.5 and cuts pricing by 20%.... description: Anthropic released Claude Opus 5.5, cutting API pricing 20% (to $4/$20 per million input/output tokens) while claiming ~40% overall savings due to fewer tokens... canonical: https://daily.dev/posts/anthropic-releases-opus-5-5-and-cuts-pricing-by-20-your-agent-calls-might-secretly-get-routed-to-a-jrjahdiam twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Anthropic releases Opus 5.5 and cuts pricing by 20%. Your agent calls might secretly get routed to an older model. | daily.dev og:description: Anthropic released Claude Opus 5.5, cutting API pricing 20% (to $4/$20 per million input/output tokens) while claiming ~40% overall savings due to fewer tokens... og:url: https://daily.dev/posts/anthropic-releases-opus-5-5-and-cuts-pricing-by-20-your-agent-calls-might-secretly-get-routed-to-a-jrjahdiam og:image: https://api.daily.dev/og/posts/jrJAhdIAM.png og:image:alt: Anthropic releases Opus 5.5 and cuts pricing by 20%. Your agent calls might secretly get routed to an older model. 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. # Anthropic releases Opus 5.5 and cuts pricing by 20%. Your agent calls might secretly get routed to an older model. **[The New Stack](https://daily.dev/sources/newstack)** · 6 min read · 1 upvotes · 0 comments ## Summary Anthropic released Claude Opus 5.5, cutting API pricing 20% (to $4/$20 per million input/output tokens) while claiming ~40% overall savings due to fewer tokens needed and 30% faster output. It beats Fable 5.1 on several agentic coding benchmarks (Terminal-Bench, FrontierCode, CursorBench) at a fraction of the cost, and enterprise testers (Box, GitHub, Deloitte) reported fewer tokens and steps needed to complete tasks. A key architectural caveat: Anthropic's safety classifiers can silently reroute individual requests within a multi-turn agent workflow to older models (Opus 4.8 or Opus 5) rather than Opus 5.5, creating inconsistency that evals assuming a single model may miss. Opus 5.5 also posted the strongest internal alignment results to date, and Sonnet 5.5 and Haiku 5.5 are expected soon. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How much does Claude Opus 5.5 cost compared to Opus 5? Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, 20% less than Opus 5's $5 and $25. Cache reads drop to $0.20 from $0.50, and cache writes fall to $5 from $6.25. Anthropic estimates overall savings closer to 40% because Opus 5.5 uses fewer tokens per task and generates output over 30% faster. _Teams comparing model costs for agentic coding can follow pricing shifts like this on daily.dev._ ### Can my Claude Opus 5.5 API requests get silently handled by a different, older model? Yes. Anthropic's safety classifiers can reroute individual requests within a workflow to Opus 4.8 or Opus 5 instead of Opus 5.5, depending on whether safeguards intervene. In a multi-turn agent workflow this means different steps could be handled by models with different capabilities, an inconsistency that evals built on a single-model assumption may not catch. _Developers architecting multi-step agent workflows track routing quirks like this on daily.dev._ ### How does Claude Opus 5.5 perform against GPT-6 Astra and GPT-5.6 Sol on coding benchmarks? On FrontierCode at default effort, Opus 5.5 beats GPT-6 Astra at roughly 20% of the per-task cost. On CursorBench 4.0, it scores 57.8% versus GPT-5.6 Sol's 41.7%, at about a third of the cost. Opus 5.5 also scored 66.4% on Terminal-Bench 4.0 versus Fable 5.1's 55.8%. _Engineers picking a model for agentic coding weigh benchmarks like these before committing on daily.dev._ --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#anthropic](https://daily.dev/tags/anthropic) [View this post on daily.dev](https://daily.dev/posts/anthropic-releases-opus-5-5-and-cuts-pricing-by-20-your-agent-calls-might-secretly-get-routed-to-a-jrjahdiam) ```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":"Anthropic releases Opus 5.5 and cuts pricing by 20%. Your agent calls might secretly get routed to an older model.","url":"https://daily.dev/posts/anthropic-releases-opus-5-5-and-cuts-pricing-by-20-your-agent-calls-might-secretly-get-routed-to-a-jrjahdiam","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/anthropic-releases-opus-5-5-and-cuts-pricing-by-20-your-agent-calls-might-secretly-get-routed-to-a-jrjahdiam"},"datePublished":"2026-09-22T16:39:46.024Z","dateModified":"2026-09-22T17:11:27.567Z","description":"Anthropic released Claude Opus 5.5, cutting API pricing 20% (to $4/$20 per million input/output tokens) while claiming ~40% overall savings due to fewer tokens...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5eb222ce50e1a86d31b3747ea9fad44a?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5eb222ce50e1a86d31b3747ea9fad44a?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The New Stack","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":"The New Stack","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/newstack","url":"https://daily.dev/sources/newstack"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/anthropic-releases-opus-5-5-and-cuts-pricing-by-20-your-agent-calls-might-secretly-get-routed-to-a-jrjahdiam","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,ai-agents,anthropic","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The New Stack","item":"https://daily.dev/sources/newstack"},{"@type":"ListItem","position":3,"name":"Anthropic releases Opus 5.5 and cuts pricing by 20%. Your agent calls might secretly get routed to an older model."}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/anthropic-releases-opus-5-5-and-cuts-pricing-by-20-your-agent-calls-might-secretly-get-routed-to-a-jrjahdiam#faq","mainEntity":[{"@type":"Question","name":"How much does Claude Opus 5.5 cost compared to Opus 5?","acceptedAnswer":{"@type":"Answer","text":"Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, 20% less than Opus 5's $5 and $25. Cache reads drop to $0.20 from $0.50, and cache writes fall to $5 from $6.25. Anthropic estimates overall savings closer to 40% because Opus 5.5 uses fewer tokens per task and generates output over 30% faster. Teams comparing model costs for agentic coding can follow pricing shifts like this on daily.dev."}},{"@type":"Question","name":"Can my Claude Opus 5.5 API requests get silently handled by a different, older model?","acceptedAnswer":{"@type":"Answer","text":"Yes. Anthropic's safety classifiers can reroute individual requests within a workflow to Opus 4.8 or Opus 5 instead of Opus 5.5, depending on whether safeguards intervene. In a multi-turn agent workflow this means different steps could be handled by models with different capabilities, an inconsistency that evals built on a single-model assumption may not catch. Developers architecting multi-step agent workflows track routing quirks like this on daily.dev."}},{"@type":"Question","name":"How does Claude Opus 5.5 perform against GPT-6 Astra and GPT-5.6 Sol on coding benchmarks?","acceptedAnswer":{"@type":"Answer","text":"On FrontierCode at default effort, Opus 5.5 beats GPT-6 Astra at roughly 20% of the per-task cost. On CursorBench 4.0, it scores 57.8% versus GPT-5.6 Sol's 41.7%, at about a third of the cost. Opus 5.5 also scored 66.4% on Terminal-Bench 4.0 versus Fable 5.1's 55.8%. Engineers picking a model for agentic coding weigh benchmarks like these before committing on daily.dev."}}]} ```