conv.

All stories
AIActive today · day 2

Anthropic launches Claude Opus 5.5, cutting price 40% while lifting coding performance

The first model in Anthropic's new 5.5 family matches Claude Fable 5.1 on most tasks at a fraction of the cost, but new safeguards and agent-breaking changes spark mixed reactions.

newest coverage 2h agosummary written 1h ago189 pieces13k likes & upvotes1.8k comments XHacker NewsMastodonNewswiresRedditBlueskyGoogle News

Part of a larger narrative

OpenAI Trade Secrets Theft

4 stories · since Sep 14 · newest 8m ago — OpenAI faces mounting legal and regulatory pressure over allegations that its engineers misappropriated proprietary technology and trade secrets from major…

  1. Anthropic says Claude autonomously discovered a novel CRISPR-like enzyme system
  2. Anthropic launches Claude Opus 5.5, cutting price 40% while lifting coding performanceyou are here
  3. OpenAI and Anthropic both ship cheaper models amid GPT-6 Astra fallout
  4. Anthropic taps Accenture for AI safety evaluations ahead of November IPO

What to know

  • Opus 5.5 is Anthropic's first model since it pledged to "pace the frontier," and the company calls it the strongest-performing model on its own alignment audit to date.
  • Pricing dropped roughly 40% for typical workloads ($4/$20 per million input/output tokens, cache reads down 60%), while output speed rose over 30%.
  • Because Opus 5.5 matches Mythos 5.1 on biology and cybersecurity capability, Anthropic applied Fable-level safeguards, and several users report losing previously approved cyber access after upgrading.
  • Reports surfaced that the release broke agent-dependent workflows and could silently route calls to an older model, tempering the otherwise positive reception.

The dispute Whether the new safety safeguards on Opus 5.5 reflect responsible caution or an overreach that degrades usefulness and trust. · positions read across 100 posts and comments

most voices

Opus 5.5 is a genuine, welcome improvement in coding quality, communication, and cost.

  • “Better than Fable, cheaper than even the last Opus. I use Opus as my main driver so this is very exciting!”

    joshstrange · Hacker News ↗
many voices

Anthropic's release cycle is a marketing pattern of hype followed by quiet quality degradation.

  • “Opus 5.5 will be great for 2-3 weeks than the nerfing starts. They give out some extra credits. It keeps degrading Opus until it's unusable.”

    gtirloni · Hacker News ↗
some voices

New safety safeguards restrict legitimate use and reduce trust in the model.

  • “Just a warning that you'll lose your Opus5 cyber access if you upgrade to 5.5 which refuses cyber capability even if you're approved.”

    mmaunder · Hacker News ↗

Anthropic Developer of Claude Opus 5.5Dario AmodeiDario Amodei Anthropic CEOMETR / Frontier Design External evaluatorsClaude Code developers Users of Anthropic's coding agent

Anthropic launches Claude Opus 5.5, cutting price 40% while lifting coding performance
x.com

How it unfolded 7 developments, newest first · click a bar or a number to jump articlespostscomments

Peak 119 pieces in one hour at Yesterday, 12 PM; 189 pieces over 2 days (74 articles · 58 posts · 57 comments) Sep 21, 10 PM — 1 piece · 1 post — X 1Sep 21, 11 PM — quietYesterday, 12 AM — quietYesterday, 1 AM — quietYesterday, 2 AM — quietYesterday, 3 AM — quietYesterday, 4 AM — quietYesterday, 5 AM — quietYesterday, 6 AM — quietYesterday, 7 AM — quietYesterday, 8 AM — quietYesterday, 9 AM — quietYesterday, 10 AM — quietYesterday, 11 AM — quietYesterday, 12 PM — 119 pieces · 62 articles · 41 posts · 16 comments — Newswires 51, X 30, Hacker News 21, +2 moreYesterday, 1 PM — 20 pieces · 6 articles · 9 posts · 5 comments — Hacker News 5, Newswires 5, Mastodon 4, +4 moreYesterday, 2 PM — 6 pieces · 1 article · 1 post · 4 comments — Hacker News 4, Mastodon 1, Newswires 1Yesterday, 3 PM — 11 pieces · 3 articles · 2 posts · 6 comments — Hacker News 6, Newswires 3, X 2Yesterday, 4 PM — 4 pieces · 1 article · 1 post · 2 comments — Hacker News 3, Newswires 1Yesterday, 5 PM — 3 pieces · 3 comments — Hacker News 3Yesterday, 6 PM — quietYesterday, 7 PM — 1 piece · 1 comment — Hacker News 1Yesterday, 8 PM — 2 pieces · 2 comments — Hacker News 2Yesterday, 9 PM — 2 pieces · 2 comments — Hacker News 2Yesterday, 10 PM — 1 piece · 1 comment — Hacker News 1Yesterday, 11 PM — quietToday, 12 AM — 1 piece · 1 comment — Hacker News 1Today, 1 AM — 1 piece · 1 comment — Hacker News 1Today, 2 AM — 2 pieces · 2 comments — Hacker News 2Today, 3 AM — quietToday, 4 AM — quietToday, 5 AM — 3 pieces · 3 comments — Hacker News 3Today, 6 AM — quietToday, 7 AM — 3 pieces · 1 article · 2 comments — Hacker News 2, Newswires 1Today, 8 AM — quietToday, 9 AM — quietToday, 10 AM — quietToday, 11 AM — 2 pieces · 2 posts — Hacker News 1, Mastodon 1Today, 12 PM — quietToday, 1 PM — 2 pieces · 2 comments — Hacker News 2Today, 2 PM — quietToday, 3 PM — 2 pieces · 2 comments — Hacker News 2Today, 4 PM — quietToday, 5 PM — 1 piece · 1 comment — Hacker News 1Today, 6 PM — 1 piece · 1 post — Hacker News 1Today, 7 PM — quietToday, 8 PM — 1 piece · 1 comment — Hacker News 1Today, 9 PM — quietToday, 10 PM — quiet 12–4567
yesterdaytodaynow · 11:11 PM ET
  1. 7

    Community debates whether gains are real or a recurring hype cycle

    Independent benchmarks (vibe-coding tests, CAD modeling, pelican-drawing tests) showed Opus 5.5 competitive with or ahead of Fable 5.1, while some commenters argued Anthropic's release cadence follows a predictable pattern of hype followed by quiet degradation.

    “Opus 5.5 will be great for 2-3 weeks than the nerfing starts.”
    — gtirloni
    1. 2 outlets Anthropic tests Fable 5.2 and Opus 5.5 ahead of the release

      first by TestingCatalog AI News, 1d ago · also The New Stack

      1 more headline

    • I have to say that I'm very pleased with this version.My prime anecdotal reason: I asked Opus 5.5 to make a sweeping change with a few Sonnet agents. It clarified the scope, we agreed and then it started. I saw some problems on the way and asked it. It's answer: This didn't go as planned. The Sonnet agents don't have the right judgement capacity…

      hgoHacker News7h agoview on Hacker News ↗
    2 more of the top 3 · 7 posts in this stretch

    • Anthropic launches new update to Claude – after telling the world to slow down AI https://www. the-independent.com/tech/claud e-anthropic-opus-5-model-update-new-b3055221.html?utm_source=flipboard&utm_medium=activitypub Posted into Tech News @ tech-news-Independent

      Independent@flipboard.comMastodon11h agoview on Mastodon ↗

    • We tested this on our agentic CAD modeling harness. Opus 5.5 seems to be a pretty dramatic improvement in spacial reasoning and CAD coding. Our go-to is still Gemini Flash 3.8 for speed and cost reasons, but Opus 5.5 is a big update.

      Mazer23Hacker News2h agoview on Hacker News ↗
    all of them →
  2. 6

    The New Stack reports Opus 5.5 broke agent-dependent workflows

    A New Stack investigation found that despite the price cut, the Opus 5.5 rollout disrupted several things developer agents relied on, and separately warned that agent calls could be silently routed to an older model.

    “Opus 5.5 will be great for 2-3 weeks than the nerfing starts. They give out some extra credits. It keeps degrading Opus until it's unusable.”
    — gtirloni, HN commenter · source
    1. 1 outlet Anthropic made Opus 5.5 cheaper. Then it broke four things your agent depends on.

      first by The New Stack, 15h ago

    • A year ago I was trying to have these models write css for me to support mobile styles for an ERP system. There were so many quirks that you typically have with css, and it couldn't figure out how to properly get it done. I'd ask it to fix something and it'd break something else - pretty much as good as any average developer is with css.Out of…

      robeymHacker News15h agoview on Hacker News ↗
    2 more of the top 3 · 9 posts in this stretch

    • > Because Opus 5.5 is comparable to Claude Mythos 5.1 in biology and cybersecurity, we’re deploying it with safeguards similar to those on Claude Fable 5.1According to their ToS, conversations flagged by the guardrails when using Fable are stored longer for moderation. That included Fable's guardrail for AI research. Moderation can mean a human…

      puttycatHacker News17h agoview on Hacker News ↗

    • The hype/degrade/hype marketing cycle that Anthropic uses to cycle between Fable/Opus or Opus/Sonnet releases is too obvious at this point.Opus 5.5 will be great for 2-3 weeks than the nerfing starts. They give out some extra credits. It keeps degrading Opus until it's unusable . By that time they are ready to release Fable 6. Which also is great…

      gtirloniHacker News21h agoview on Hacker News ↗
    all of them →
  3. 5

    Early users report strong coding gains alongside new restrictions

    Developers testing Opus 5.5 in Claude Code and third-party tools praised its speed, clarity and coding output, while others flagged that upgrading could strip away previously approved cybersecurity access.

    “Just a warning that you'll lose your Opus5 cyber access if you upgrade to 5.5 which refuses cyber capability even if you're approved.”
    — mmaunder

    • I've asked Claude Opus 5.5 to create 100 HTML files Rules were simple: - look stunning - zero repetitive designs - go full creative mode CLAUDE OPUS 5.5 IS THE BEST MODEL I HAVE EVER TESTED !!! The previous best results on this experiment was of Fable 5.1, but Opus 5.5 is NEXT

      @MiaAI_labX1d ago38▲view on X ↗
    2 more of the top 3 · 22 posts in this stretch

    • Opus 5.5 is neck-and-neck with Fable 5.1 and Astra 6 in my vibe-coding tests - maybe even better than Fable 5.1Minecraft clone: https://senko.net/vibecode-bench/2026/voxel-opus-5.5.html (Opus 5.5) vs https://senko.net/vibecode-bench/2026/voxel-fable-5.1.html (Fable 5.1) vs https://senko.net/vibecode-bench/2026/voxel-gpt-6-astra.html (Astra…

      senkoHacker News1d agoview on Hacker News ↗

    • Initial usage of Opus 5.5, much better than Opus 5, you can actually understand what it's saying, it actually executes without second guessing itself, and is pretty quick and thorough on implementation. Opus 5 was one of the worst models that Anthropic ever produced imo, 5.5

      @HackingDaveX1d ago13▲view on X ↗
    all of them →
  4. 4

    Press outlets detail pricing, safety and IPO context

    Outlets including The Verge, MacRumors, Bloomberg, TechCrunch and ZDNET covered the release, noting new cybersecurity/biology safeguards similar to those on Fable 5.1 and framing the release against Anthropic's reported IPO plans.

    1. 6 outlets Anthropic introduces Claude Opus 5.5, its new lower-cost AI model for agentic coding, computer use, and more

      first by Business Today, 1d ago · also Reuters, Neowin, Bloomberg, 9to5Mac, Investing.com News

      5 more headlines

    • Opus 5.5 performs at the level of Fable 5.1. It's ~30% faster and ~40% cheaper than Opus 5 per task. In Claude Code: - 5-hour session limits increase 20% today - Opus 5.5 is priced lower, so it goes 25% further within limits - Pro, Max, and Team users get a reset to use anytime

      @ClaudeDevsX1d ago2.5k▲view on X ↗
    2 more of the top 3 · 58 posts in this stretch

    • Here are pelicans for thinking levels low, medium, high, and xhigh: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...All four levels have a correctly shaped bicycle frame. The differences between the pelicans aren't huge, but the xhigh one has a better beak.I haven't managed to get one for level "max" yet, it hit the limit of 128,000…

      simonwHacker News1d agoview on Hacker News ↗
    • Techmeme@techhub.social

      Anthropic raises its five-hour usage limits by 20% on Pro, Max, and Team plans, and gives subscription users a rate limit reset (David Gewirtz/ZDNET) https://www. zdnet.com/innovation/anthropic -claude-opus-5-5-fable-5-1-performance-costs-less/ http://www. techmeme.com/260922/p41#a26092 2p41

      Techmeme@techhub.socialMastodon1d agoview on Mastodon ↗
    all of them →
  5. 3

    Anthropic and Claude accounts confirm same-day availability

    Official Anthropic and Claude social accounts announced the model was live immediately, with Claude Code getting a 20% increase in five-hour session limits and a one-time rate-limit reset for subscribers.

    “Opus 5.5 performs at the level of Fable 5.1. It's ~30% faster and ~40% cheaper than Opus 5 per task.”
    — @ClaudeDevs

    • Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.

      @claudeaiX1d ago5.6k▲view on X ↗
    2 more of the top 3 · 3 posts in this stretch
    • TechCrunch@mstdn.social

      Anthropic called it "the strongest-performing model we've tested to date." https:// techcrunch.com/2026/09/22/anth ropic-releases-opus-5-5-with-lower-prices-and-fable-level-performance/?utm_source=dlvr.it&utm_medium=mastodon

      TechCrunch@mstdn.socialMastodon1d agoview on Mastodon ↗

    • Claude Opus 5.5 is available today.

      @AnthropicAIX1d ago1.3k▲view on X ↗
    all of them →
  6. 2

    Anthropic launches Claude Opus 5.5

    Anthropic released Opus 5.5, the first model in its new 5.5 family, saying it performs at the level of Claude Fable 5.1 on most work while costing 40% less to run than Opus 5, and calling it the strongest-performing model on its automated behavioral audit to date.

    “Anthropic called it "the strongest-performing model we've tested to date."”
    — Anthropic

    1. 13 outlets first by TestingCatalog AI News, 1d ago · also Digital Trends, Pulse 2.0, iClarified, Constellation Research, Thurrott, RuntimeWire +6 · read ↗

    2. 12 outlets Introducing Claude Opus 5.5

      first by Anthropic, 1d ago · also Digit, Moneycontrol, Tech in Asia, Lenny's Newsletter, VentureBeat, Decrypt +5

      9 more headlines
    7 more claims →
  7. 1

    Signs of an imminent Opus 5.5 release surface online

    AI-tracking accounts spotted preparation artifacts suggesting a new Claude Opus release was imminent, ahead of the official announcement.

    “Early preparations for Claude Opus 5.5 have been spotted again. Seems imminent.”
    — @M1Astra

    • Early preparations for Claude Opus 5.5 have been spotted again. Seems imminent.

      @M1AstraX2d ago220▲view on X ↗

What people are saying 11 voices from 4 sites · best of 100 · verbatim

Still unanswered
  • Will Opus 5.5's quality quietly degrade over the coming weeks as some commenters predict happened with prior releases?
  • Does upgrading to Opus 5.5 permanently strip previously granted cybersecurity/biology access, and is there a reliable workaround?