Anthropic reports elevated errors across multiple Claude models
A status-page incident coincides with commenter reports of overly sensitive safety flags and rumors of an imminent Opus 5.5 launch.
What to know
- Anthropic's status page logged elevated errors for multiple Claude models overnight into September 22, 2026.
- Commenters report the errors coincide with, and may be caused by, an imminent Opus 5.5 launch that has been rumored for days.
- A separate complaint surfaced in the same thread: unusually aggressive 'flagged by safeguards' triggers forcing automatic downgrades from Opus to Sonnet during normal use.
- Some developers describe routing around the outage by handing off in-progress Claude sessions to other coding agents like Codex.
The dispute Whether the outage stems from load/rollout issues around an imminent Opus 5.5 launch or from an unrelated tightening of safety-filter sensitivity. · positions read across 19 posts and comments
The outage is likely tied to an imminent Opus 5.5 release, following a pattern of incidents preceding launches.
-
“There are rumors of Opus 5.5 launching on Tuesday so maybe it due to that? Previously as well there have been incidents just before launch.”
makeavish · Hacker News ↗
Safety safeguards have become overly sensitive recently, disrupting legitimate work.
-
“I have been getting a lot of stuff "flagged by safeguards" in the last two days... If anyone here works at Anthropic, did you make the safeguards super sensitive recently?”
tombert · Hacker News ↗
When Claude is unavailable, workflows can continue by switching to other agents/harnesses without much friction.
-
“I was in the middle of work with Claude could you read the plan and the session "x" in ~/.claude and continue the work" it just completed everything :D”
mariocesar · Hacker News ↗
There is irony in an AI touted as replacing developers struggling to maintain basic reliability.
-
“Claude is both on the verge of replacing all software devs and keeping a two 9 SLA”
theophilus76 · Hacker News ↗
Anthropic Operator of Claude and its status page
How it unfolded 4 developments, newest first · click a bar or a number to jump articlespostscomments
-
4
User reports unusually aggressive safety-filter triggers
A commenter reported a sharp rise over the prior two days in requests being 'flagged by safeguards,' forcing automatic downgrades from Opus to Sonnet during ordinary reverse-engineering work.
“I am quite confident that what I'm doing is well within the law... but I can't even use Fable anymore because every time I enable it, it works for about twenty seconds and makes me drop down to Opus 4.8, and often even down to Sonnet.”
— tombert -
I don't know if something changed recently, but I have been getting a lot of stuff "flagged by safeguards" in the last two days.I am quite confident that what I'm doing is well within the law, and I'm not even doing any kind of pen-testing stuff, just some basic reverse engineering, but I can't even use Fable anymore because every time I enable…
2 more of the top 3 · 16 posts in this stretch
-
I’m working on a project that makes switching between coding harnesses essentially unnoticeable. It’s particularly useful when I run out of credits on any given dayThe entire platform is skill driven, and based on the premise that state is your local file system. That makes switching harnesses so easyIt’s all open source and has plenty of other…
-
Apologies for the rant, but why do these things constantly hit the front page? It is not interesting, it's not a discussion, and if you are using the models you probably already know.HN is already a waterfall of AI meta conversations and bike-shedding, now we have to discuss service outages about the AI too?Can we talk about stuff people are…
-
-
3
Users describe workarounds by switching to other coding agents
Some commenters described routing around the Claude disruption by handing off in-progress work to alternative tools like Codex or custom harness-switching setups.
“I was in the middle of work with Claude could you read the plan and the session "x" in ~/.claude and continue the work" it just completed everything :D…”
— mariocesar -
I'm was closing a day of work and it stopped in the middle of a long plan, I started codex and tell it: "I was in the middle of work with Claude could you read the plan and the session "x" in ~/.claude and continue the work" it just completed everything :D
-
-
2
Commenters link outage to rumored Opus 5.5 launch
Several Hacker News commenters speculated the errors are tied to a rumored imminent release of Opus 5.5, noting that past incidents have preceded major launches.
“There are rumors of Opus 5.5 launching on Tuesday so maybe it due to that? Previously as well there have been incidents just before launch.”
— makeavish -
There are rumors of Opus 5.5 launching on Tuesday so maybe it due to that? Previously as well there have been incidents just before launch.
-
-
1
Anthropic posts incident for elevated Claude model errors
Claude's status page reported an incident of elevated errors affecting multiple models, which was picked up and shared on Hacker News and Mastodon.
-
first by HN Frontpage, 2d ago
-
What people are saying 12 voices from 1 site · best of 19 · verbatim
- Did Anthropic actually increase safeguard sensitivity in the past two days, and if so why?
- Is the elevated-error incident directly caused by preparations for an Opus 5.5 launch?
- Does hitting cached-pricing errors during the incident cause users to lose expected cached-price discounts?
- Sep 22
-
This coincides with the release of Opus 5.5. Wonder if that has any causal relationship to this?
-
They (Claude, OpenAI, and others) does outstages to "see how we are important?" - and then release a new model and nerf all others
-
omp can switch seamlessly between models in the same session. I really like this feature.
-
I don’t understand your point or why people are upvoting this. I’ve done this between many different models. Did you just discover that a frontier model can read context? I genuinely don’t get your point. I’ve had to have Claude agents pick up codex’s context after it hits one of its “cannot connect” issues way more often than the other way around.
-
I have had the same experience. Switching tools is much easier when the project notes are clear and up to date.
-
Both Cursor and Grok build ask if you want to resume a session from another harness. Quite brilliant
-
I don't believe we are anywhere close to AGI until I see 99.95% uptime at the frontier AI lab. Now they are undercounting. I hit 50x even when they claim they are up.We now have alien intelligence that is very different from human intelligence.
-
Related, have found this handoff skill invaluable for moving tasks between harnesses:
-
I have a deja vu like this happens everytime before they launch a new family of models
- Sep 21
-
Opus 5.5 and fable 5.2 will release tonight.gpt-6-sol and Aeon (personal agent) on Thursday. Already preceded by a huge week with step, mimo, grok, and jev releases.Relentless cycle.
-
When this happens, how often does it cause us to lose our cached pricing? (when we should be getting the cached price)
-
Claude is both on the verge of replacing all software devs and keeping a two 9 SLA