Opus 5.5 feels way dumber after today’s outage, anyone else noticing this?
Users are reporting a significant decrease in the performance of Opus 5.5 following a recent outage. One user, who previously found Opus 5.5 impressive for legal work, writing Minecraft mod scripts, and troubleshooting NAS setups, noted that after Claude went down and the service returned, Opus 5.5 felt "noticeably, almost ridiculously stupid," like a "quantized version or a completely different model." The user emphasized that the drop in quality has been "really really really really noticeable."
Time & source
Times shown in UTC
Display time zone: UTC
Local time zone unavailable; showing UTC.
PublishedOffset at this time: UTC+0Sep 30, 2026, 00:57 UTC
IngestedOffset at this time: UTC+0Sep 30, 2026, 08:00 UTC
- Published
- Sep 30, 2026, 00:57
- Ingested
- Sep 30, 2026, 08:00
- Source type
- Dev community
- Tier
- Community
- Source status
- Healthy
Tier is a per-source editorial setting, not a per-item score.
Anyone else feel like Opus 5.5 suddenly got way dumber after today’s outage?
I’ve been using Opus 5.5 for my legal work and it had honestly been impressing me, also for writing scripts for Minecraft mods and helping me understand my NAS setup and troubleshoot bugs, but today, after Claude went down and the service came back, it suddenly felt noticeably, almost ridiculously stupid, like a quantized version or a completely different model, it just doesn’t feel like the Opus I was using before
Now it forgets things, makes mistakes, invents details that weren’t even in the documents I gave it, makes typos and needs way more back and forth to finish the same kind of task, with my Minecraft mods it started introducing bugs, then constantly apologizing and giving me the usual “you’re right” response instead of actually fixing the problem, that kind of excessive agreement was something I mostly experienced with Gemini before
With my NAS it suddenly started forgetting the context and how everything connects, my setup involves Tailscale, firewall rules, the router and syncing with another NAS, all of that matters when we’re trying to figure out a bug, we’d been working through this all week without any problems, then today it forgot something as basic as the fact that the connection goes through Tailscale and started giving me advice as if I had a standard local network setup, completely overlooking the architecture we’d already been working with
That’s what makes the difference so obvious to me, I’m continuing work it was handling well before the outage, and suddenly I have to keep reminding it of basic details, correcting things it made up and going through several more attempts to get something usable
Anyone else experiencing this since the service came back, because the drop in quality today has been really really really really noticeable on my end