YHK - yet another hacker news clone

Is Opus 5.5 nerfed? Thinking patterns changed on November 3

105 points rzk 2 hours ago 58 comments

Retro_Dev 2 hours ago | parent

As one user pointed out:

> I noticed this too as did many others. Had to bump my default reasoning but it’s not the same. Seems to often happen before new model drops

I heard that Qwen4 was dropping some time soon, I wonder if Anthropic is trying to announce a new model immediately after Qwen4 - both for their IPO, and to kill the competition.

Gareth321 1 hour ago | parent

This has happened to consistently across OpenAI and Anthropic models that I can't believe people are still denying it. Obviously they're silently quantising their models or making other efficiency changes. If they can save 40% RAM and only hurt intelligence by a few percent in their benchmarks, they're going to do that. Unfortunately their benchmarks are obviously not representative of the real world, and these downgrades are very noticeable for people who rely on these models for work.

Of all the reasons to look at local inference, this is the one which pushes me in that direction the most. I want to rely on a model for consistency in my workflows. If I wake up on Monday and suddenly it's refactoring things it built on Friday, that's a huge problem. That makes it unreliable.

fastball 1 hour ago | parent

I guess the question really is: is this malice (trying to trick people into thinking new model is better by nerfing old one), or is this capacity constraints, where they obviously need to make room in their infra for inference of the new model by shoving the other models into a smaller box?

timschmidt 58 minutes ago | parent

Little of column A, little of column B.

grey-area 1 hour ago | parent

These machines are inherently unreliable.

yfontana 54 minutes ago | parent

> I can't believe people are still denying it

Probably because those "nerfs" do not show on sites that track model performance over time. Sometimes Anthropic/OpenAI mess stuff up in their harness, but usually not in the models themselves.

xscott 29 minutes ago | parent

It's easy to imagine they can detect when they're being profiled and route accordingly. Paranoia or not, it's possible.

terminalcommand 1 hour ago | parent

I am using Opus 5.5 high almost exclusively. I definitely felt this and looked for nerfing watch sites. It felt like downgrading to the old version of deepseek. Things have improved a bit now.

sunaookami 1 hour ago | parent

It's just a new summarizer, the model and performance itself is unchanged.

kolinko 1 hour ago | parent

+1, this tells nothing about the quality. Or even whether this is an upgrade or downgrade.

xscott 1 hour ago | parent

With open weights, YOU get to decide when the model changes.

beebmam 1 hour ago | parent

Not true at all, there's much more to a model than its weights.

digitalPhonix 1 hour ago | parent

Like?

phyalow 1 hour ago | parent

Thinking budgets…

digitalPhonix 1 hour ago | parent

In the context of this question, that's still something you as the user control.

Specifically

> YOU get to decide when the model changes

xyzsparetimexyz 1 hour ago | parent

Yeah obviously but you can still control it.

f6v 1 hour ago | parent

You can still control it if you can trivially spin up a node with eight B200s and 2TB of memory to run your Deepseek v4 instance. Nobody on a subsidized plans is doing that. Local models that fit 128GB memory don't come close even to a "nerfed" frontier model.

Roark66 25 minutes ago | parent

Actually Qwen3.8-flash-next comes very close to a "nerfed frontier model" the way I host it locally at 4.75bpw with int8 kv. It is pretty much equivalent to opus 4.6 up to 4.8 on coding. What it lacks is some factual knowledge - this is to be expected, it's a far smaller model. But if you run the kind of coding tasks where you have to spec them in high detail anyway (large very custom environment) this model is very usable.

I run it on a 6*rtx3090 consumer PC (no threadripper etc sadly) with very constrained pcie topology (4 cards sit on 2 lanes basically) and i still get 7k prefill and 50t/s single user or 200t/s 7 users. All at full 262k contexts of course and with 96GB of ram kv cache that means I can easily have 10+ Claude code sessions open with 1-7 generating at any time.

I've been the person that said, before opus 4.6 came out that this AI stuff is absolute crap for programming. Then it suddenly wasn't and I was using it many hours each day.

Local AI was the same. It was all crap. Models like GLM5.3 Flash that are many times larger than that qwen, lots of people are oooing and ahhing over I found way too inferior. Previous DeepSeek V4.1 Flash 0731 was atrocious. They benchmaxed it to hell, but at least it was small.

The current model of same name but release 1017 and double the size is currently the best open source model IMO. It even beats my self hosted Qwen3.8-flash-next.

I've tried hosting that DeepSeek model, but even adding another rtx3090 for a total of 7 cards and my 192gb of ram bought when it cost under $1000 I get only about 25t/s decode and few hundred prefill. That is far too slow for me to use.

But, I believe once this pyramid scheme of circular financing finally bursts there will be a glut of hardware on the market where people like me will be able to buy equipment to run 750B models like we now run 120B.

Sphax 15 minutes ago | parent

Just curious, did you put all these cards in a server or desktop case ? and what kind of motherboard ?

Also, how much power does that consume at full tilt ?

Tepix 20 minutes ago | parent

For Deepseek V4 Flash (not 4.1), you can use two 128GB unified memory boxes and run it at the official weights (mostly 4bit) with lots of context.

That's around 1% of the price of 8x B200. At current prices, it's as low as $6200. At last year's prices that's $3700 or so.

Giorgi 1 hour ago | parent

Yes, I noticed it too, dumbing down continues, there goes their IPO.

hebelehubele 1 hour ago | parent

They're preparing to release a new model, so it has to look better than the current one.

ghoul2 1 hour ago | parent

Definitely a significant change. I noticed it in just daily work. Quality of work down, mistakes up, repetitive errors in scripts - rare earlier, now a few every session.

listedb 1 hour ago | parent

It's dumbing down a little bit but I don't think its a lot worse. I think the change in display adds to the feel

tammyselvi 1 hour ago | parent

not as bad as sol 6.1. im getting better results out of glm 5.3

cromka 1 hour ago | parent

I was down voted into oblivion when I said that 3 weeks ago. Both Codex and Claude are absolutely nothing compared to what it was.

benwills 1 hour ago | parent

I noticed Fable decline in quality a couple weeks ago. And I've noticed Opus drop over the last week, even after bumping to a default of Extra High reasoning for almost everything.

Tonight has been maddening with the kinds of absurd mistakes it's repeatedly making. It's so frustrating.

I'm the kind of person that wants something to be consistent and reliable. And Anthropic has been great, except for in this way. Which is enough of a recurring frustration (about every couple months this year) that I genuinely can't wait for other options to catch up. Then I'll be leaving as soon as I can.

I just don't know how long that will be.

jdcasale 1 hour ago | parent

I've been using Fable for a couple months as a relatively independent, long-horizon task executor and it worked well. About two weeks ago it completely stopped working. Like ~85% success rate down to ~10% in my workflows.

Totally possible that I'm just hallucinating/projecting/etc or my workflows somehow changed in terms of difficulty, but it's been kind of shocking.

cromka 1 hour ago | parent

I have the same exact experience. Working on one project, brainstorming. It was brilliant in July and August, like a seriously smarter, more experienced version of myself, reading in my mind, one step ahead of me, when pointed out a mistake it would remedy on the spot with the exact brilliant solution.

Now it's like talking to my hungover self, after 3 days bender, who just wanted to be left alone.

Roark66 40 minutes ago | parent

Same thing for me, but the day Fable got lobotomised Opus 5.5 started actually working pretty well for me.

shikck200 1 hour ago | parent

Token use exploded a week or two ago. I worked for a few hours and hit my 5 hour limit. Did only some simple bug fix. Somethings definitely broken.

This is the #1 thing i HATE with LLMs. They are not predictable like a compiler is.

cromka 1 hour ago | parent

It's not LLMs! Its their providers! If you self-hosted, you'd observe the slow down but you'd get same quality.

KeplerBoy 37 minutes ago | parent

It's probably not token usage, but a flexible 5 hour limit. There's a reason why they only state it in percentages.

monster_truck 27 minutes ago | parent

They're actually extremely predictable. What you are experiencing is the provider you are entrusting with your time, effort, and money fucking you because they need the capacity to train the next one. So without any real notice or indication other than things vaguely seeming broken in ways everyone seems to be experiencing differently, they're serving up some shitty quant of what you already paid to use.

davidd_1004 1 hour ago | parent

Seems fine to me

peri-cl 1 hour ago | parent

Eagerly awaiting Artificial Analysis to redo their Anthropic benchmarks and kick it down 15-20 index points, in the interests of accuracy. Those guys love accuracy. They moved Heaven and Earth to kick Kimi off the #2 spot, out of their deep, abiding passion for having the most accurate benchmarks in the industry.

I have no doubt they will ensure their users get an accurate picture of Opus 5.5 within hours.

MallocVoidstar 10 minutes ago | parent

If Anthropic are doing A/B testing it almost certainly only affects subscription users, not API users. Every benchmark I know of uses the API.

xbmcuser 1 hour ago | parent

From what I understand they are picking and choosing who is running the latest models. So nerfed yes but for you not for everyone

tsukikage 1 hour ago | parent

LiveNerf suggests not.

https://github.com/ninjahawk/livenerf

overflowy 54 minutes ago | parent

My paranoia tells me they A/B test users which means consistent benchmarks go through the high performance pipeline while regular requests fall back to the quantized / experimental / low cost version. I can see the main reason being the oldest one in the book, which is making more money.

musha68k 45 minutes ago | parent

It could also be a regional effect where systems are under higher load and hence dynamically routing to / provisioning lower precision instances; or some other factor where one is affected whereas someone else isn't.

eithed 34 minutes ago | parent

It's not paranoia and would align with https://www.reddit.com/r/ClaudeCode/comments/1xr9x5y/the_ner...

I wonder as well whether would this pop up on nerfbench

igleria 43 minutes ago | parent

as a user I believe it: they are probably reducing assigned compute to prepare to offer new models.

That would be the less bad reason.

lukax 41 minutes ago | parent

I guess it's just capacity problems. They probably have a fixed amount of compute and they need to make space for new models that will be released. So they quantize existing models e.g. in half so that can use 50% of current capacity for new models. They have such massive scale that you just cannot dynamically get that amount of new compute online so they need to work with what they have. No provider has thousands of GPUs idle.

PUSH_AX 41 minutes ago | parent

Fable 5.1 quality took a significant dive for me last week. I'm not imagining it.

throwaw12 39 minutes ago | parent

LLMs are such a great business, they are non-deterministic, hence you can do whatever you want, because customers can't do anything about it:

* nerf them based on supply and demand, when you have a high demand, nerf it, but keep charging same price.

* do you want to show revenue growth to investors? increase thinking tokens, customers will keep paying, because again, this is non-deterministic system, sometimes it thinks more

* do you want to increase profits? start serving 4-bit, 5-bit quantizations and keep charging same price

* want even more? publish a paper "Advanced Auto Super Smart Prompt Router - AASSPiRation", and start routing requests to Qwen3.8-27B or GemmaE2B from time to time on the edge

* Not enough? introduce "Complex Chained Workflows - CCW" in your coding harness, which spawns multiple deterministic workflows (which you can't stop or control) and they will use Haiku to write the code, then use Opus 5.5 high to verify the code (which is obviously wrong) and Opus 5.5 produces recommendations on how to fix the code, delegates to Haiku, which again fails writing proper code, and keep doing it 5-6 times, you made lots and lots of money (don't forget, all should run in parallel, to pick one best result among 10 outputs).

ricardobeat 38 minutes ago | parent

I noticed the same, and posted about it three days ago.

Opus is suddenly fumbling basic tool calls and failing to finish tasks even at high, where I had been permanently using it on low effort with success since it came out. It is very frustrating.

My guess is they quantized it and/or reallocated DC capacity to train a new model, this has happened before.

jonathanlydall 33 minutes ago | parent

Anecdotal, but I tend to get by with medium thinking on Opus and I noticed that as of sometime last week it really did feel like it had become noticeably slower to think about things.

I wouldn't say I've especially noticed its effectiveness having been reduced, it's just slower.

perks_12 31 minutes ago | parent

Welcome to the world of post-IPO AI. Now the 20% growth needs to come somewhere, and profitability would be a nice starter.

Asmod4n 30 minutes ago | parent

Hm, I’ve seen a degradation of 5.5 yesterday. I’ve had a session running for around 50 days and it just started to generate garbage and ignored all rules. A fresh one feels the same as always

etcetcetcetceta 23 minutes ago | parent

I suppose they can achieve a benchmark then nerf the model to save compute once they've got the headline.

0x70dd 22 minutes ago | parent

Ok, I'm not going crazy then. Noticed that it is slower and results were of a lower quality.

aurad 15 minutes ago | parent

I've seen big drops in Opus 5.5 (on high/x-high) & Fable. Both in terms of following instructions / workflows execution and overall code quality. Also the token burn-rate increased a lot, I've reached limits of 3 Ultra subscriptions during last week & had to spend 1.5k on API usage during the weekend just to be frustrated with the stupidity of the output.

Weirdly enough, Sonnet quality seems unchanged.

fractalf 10 minutes ago | parent

5.5 is useless below max effort, just frustrating and plain old wrong