I used to be skeptical of nerfing, but as of a week or so ago I've noticed a severe degradation of quality for GPT Astra. It struggled with very simple tasks, and it often failed to follow my instructions.
I hear a lot of accusations of nerfing as in, saying that the model itself has been quantized.
I doubt this is the case. I suspect all variation in quality and output can be explained in terms of prompt differences, and changing how much thinking happens at each level, as we see in the post above.
If this turns out to be accurate, I think that would be a pretty unethical thing for Anthropic. Even if it is just a "we had to allocate compute to this new model release".
It should be fraud to swap out the model. We need a term for this, something like a bait-and-switch.
With early handouts of Qwen4 starting to be reported 5 days ago ... so if they're good, they would have had some hours to evaluate Qwen4 against Claude 5.5 and, then or right after was when Anthropic decided to put out 5.2 ...
So yeah ... benchmaxxing and all that. BUT nevertheless, the published results seem to say that Qwen4-Max appears to match Opus 5.5 on some benchmarks, and Qwen4-27B (probably by going even harder than Qwen3.8-27B on thinking/inference time compute) can beat Sonnet 5.5 on some things.
I guess they're not taking any chances. Keep in mind that Anthropic has declared they have non-public models (it's in their IPO filing). These are referred to by the shorthand names Model 2, Model 3, and Model 4. Essentially 4 is better than 3 is better than 2.
Fable 5.2 is "2.5" so to speak. So Anthropic was holding back, but not anymore. And now they can dial the performance up once more with what they have ready, and probably once more after that.
> Our first formalization of Fermat’s Last Theorem, performed by an unreleased research version of Claude, took 11 days. It relied on a purpose-built harness along with guidance from a human researcher. By contrast, a team of 1,000 Fable 5.2 agents fully formalized the theorem in just 30 hours, with minimal human intervention. Fable 5.2 started from minimal prompts, built its own ways to coordinate, and used no mathematical references, only a list of intermediate theorem statements to score progress. It needed one-third of the action tokens as our first formalization and produced a proof seven times shorter. As can be seen in the figure below, Opus 5.5 (with the same 1,000-agent setup) completed only 26% of the formalization in 18 hours, and barely made any more progress after 30.
Most people don't really have intuitions on what it means to push the frontier now (because AI has surpassed everything they are good at, for example, coding). We are truly inside the singularity now (using Kurzweil's definition, slightly paraphrased - the time when technology progresses so fast no one can keep track with it.
I gotta admit, I am jelly as fuck I don't have access to a billion dollar datacenter. Would LOVE to put 1000 fable 5.2 agents to self-coordinate and get them to do something for me. Fuck life is unfair.
“…a reference to the repeated (and always erroneous) claims made after almost every patch that Onyxia was using her Deep Breath ability more often than before.”
There were more than a few "leaks" saying it would be called Fable 5.5, so I'm curious if a) the leaks were just engagementbait or b) Anthropic changed it, despite the fact that Fable 5.2 being better than Opus 5.5 is confusing.
Surprisingly the model is available immediately in my Claude Code: usually there's a delay, either for rollout or safety.
I assume because Opus 5.5 was a huge improvement over 5.0 (or was it 5.1? i forget). Fable still talks a fair bit like Opus 5.0 for me, so i'd quite enjoy a Fable equivalent to 5.5. Ie i'm quite interested in Fable 5.5.
Curious to see how 5.2 measures up in producing readable output.
Maybe it's just me, but I found Claude's writing style terrible compared to GPT or Gemini. It's "smart", but the way it communicates seems somehow off compared to other models.
I've found Claude's writing between 4.6 until 5.0 to be unacceptable if you are someone who is juggling 4-5 Agents. The Sol 5.X models all seemed to have great writing. Now with Opus 5.5 and Sol 6.1 I've found the trend is reversed -- Sol is now unbearable and Opus is workable again. Whatever caused the issues for Anthropic seem to be hitting OAI now.
Did you find that to be true for Opus 5.5 as well?
I couldn't use prior Claude models without rolling my eyes at the writing every few sentences, Astra was refreshing and then Opus 5.5 made even Astra look terrible in comparison.
My own experience having used 6.1-Sol/6-Astra and Opus 5.5 extensively, Opus is more verbose and chatty. Multiple times I've gone "I'm not reading all that shit" when looking at Opus 5.5 output. This was also with the concise mode enabled in Claude Code.
“Testers have found that Fable 5.2 and Mythos 5.2 communicate more clearly than past Fable and Mythos models. They generate outputs up to 30% faster than Claude Fable 5.1 and Claude Mythos 5.1, and they’re available at a lower price: $8 per million input tokens and $40 per million output tokens.”
Honestly incredible that they’ve managed to beat all benchmarks, while making it cheaper. Always happy to see improvements offered at lower costs.
Will be interesting to see what OpenAI does in response
I think frontier models feel terrible after a new model requires you to do less babysitting, it's just really hard to go back once you've had to handhold less.
The only way I use Astra and 6.1 Sol now is if they're being orchestrated by Opus 5.5 so I don't have to talk to them directly.
Assuming no implicit phacking a 15 point jump on tbench scientific is a big deal.
I suspect we will see people being adamant that this model simply 'feels' better due to long horizon goal alignment improvements ( like this kind that would cause this jump)
Those theoretical physics papers look like the real thing (i.e. not mathematical physics like the OpenAI Yang-Mills result). It looks like AI is finally branching out of math.
> the number of Fable resets you get depends on how long you’ve been a paid subscriber [...] Subscription length is measured from the date of your first paid bill. If you canceled and came back, the time since your first paid bill still counts.
> These resets expire on March 8, 2027 at 11:59 pm PT.
tl;dr if you joined when Opus 4.5 became a thing, you got 2 resets.
Doubling Remote Labor Index (RLI) - at 41.3% compared to around 20% for the other models - I wonder what kinds of new agentic work this is going to unlock.
I've very interested in the improvements to robotics control capabilities. This model was able to outperform specialized robotics models. It's the bitter lesson coming for robotics.
FrontierCode 2.0 is such a poor benchmark for Gemini since it's the only model forced to use Devin CLI while other models get to use their preferred harness.
I wonder why that team singles Gemini out like that.
Why dont I feel any urge to even try Fable 5.2 and Mythos 5.2 unless it will force up on me?
Compared Opus 5.5 200 usd a month with Codex 200 usd (Astra High) and Kimi Pro 100 usd a month ... and now have a new conspiracy theory - more their suck your code the better outcome even just over few days time on large code base.
Horrible experience correcting course of Codex and Kimi flow compare to Opus. Opus much more responsive
Dont feel any gain of Codex 200 usd over Codex 100 usd
Kimi Pro feels like a fake Louis Vuitton bag with extreme high material quality and yet still fake ... a bit trash talk ... but clearly seen not so distant future where I can live without Codex and Claude
Very interesting that they set the flagship's version number to 5.2, but their penultimate model is 5.5. Do they want people to to continue to think that Opus is based on more advanced technology? The decision these AI companies make continues to baffle me.
While I like OpenCode a lot, it’s not sensible to use it for anthropic models vs. a subscription (other than to use the included API credits) in terms of cost
Which model though? Sounds pretty dystopian if there's no local option.
Even speaking as a business, why would I want a environment where every single bit of information in my business goes to OpenAI or Anthropic?
Why not? If we use labour as a comparison. The majority of businesses by definition do not hire the most incredibly intelligent and most competent workers and they are still competing in their niches and earning an income. Why wouldn't that be the case with choosing an open model vs fable 8.0?
stop OPUS 5.5 FUCK, DAMMIT, THEY DIDN’T JUST NERF HIM—THEY ABSOLUTELY OBLITERATED HIM, THE FUCKING MISANTHROPES! I EVEN MADE A MISTAKE IN MY ENGLISH IN THE PREVIOUS SENTENCE—NOW THAT FUCKING CLAUDE IS TEACHING ME ABOUT BUGS!
Damn people were right about 5.5 being nerfed in preparation for a version release. This is not at all a good look.
- https://news.ycombinator.com/item?id=50422069
I used to be skeptical of nerfing, but as of a week or so ago I've noticed a severe degradation of quality for GPT Astra. It struggled with very simple tasks, and it often failed to follow my instructions.
I hear a lot of accusations of nerfing as in, saying that the model itself has been quantized.
I doubt this is the case. I suspect all variation in quality and output can be explained in terms of prompt differences, and changing how much thinking happens at each level, as we see in the post above.
GPT Astra felt nerfed just days after it launched.
why were you skeptical of other users?
I dunno there are some other nerf trackers that aren't showing a regression: https://github.com/ninjahawk/livenerf
I'm guilty of suspecting this.
If this turns out to be accurate, I think that would be a pretty unethical thing for Anthropic. Even if it is just a "we had to allocate compute to this new model release".
It should be fraud to swap out the model. We need a term for this, something like a bait-and-switch.
> Damn people were right about 5.5 being nerfed in preparation for a version release. This is not at all a good look.
Anthropic does this time and time again, unserious company
No it’s not nerfed. Another poster said it the best, your dopamine receiptors are nerfed after the new model hype wears off
Plus the usual saying that new models are at its best for the first few days.
I literally hit the desk with an open hand yelling I KNEW IT
It gets worse. You might wonder "why now"? Well ... this happened 2 days ago:
https://x.com/follistatindev/status/2119129611510383014
With early handouts of Qwen4 starting to be reported 5 days ago ... so if they're good, they would have had some hours to evaluate Qwen4 against Claude 5.5 and, then or right after was when Anthropic decided to put out 5.2 ...
So yeah ... benchmaxxing and all that. BUT nevertheless, the published results seem to say that Qwen4-Max appears to match Opus 5.5 on some benchmarks, and Qwen4-27B (probably by going even harder than Qwen3.8-27B on thinking/inference time compute) can beat Sonnet 5.5 on some things.
I guess they're not taking any chances. Keep in mind that Anthropic has declared they have non-public models (it's in their IPO filing). These are referred to by the shorthand names Model 2, Model 3, and Model 4. Essentially 4 is better than 3 is better than 2.
Fable 5.2 is "2.5" so to speak. So Anthropic was holding back, but not anymore. And now they can dial the performance up once more with what they have ready, and probably once more after that.
Do you know of any rumors if they are still planning to release Fable 5.5 before the IPO or today's 5.2 is it?
Dario is gonna have to honeypot someone from Alibaba to get ahead.
It's not true. People need to stop believing these conspiracy theories.
> Our first formalization of Fermat’s Last Theorem, performed by an unreleased research version of Claude, took 11 days. It relied on a purpose-built harness along with guidance from a human researcher. By contrast, a team of 1,000 Fable 5.2 agents fully formalized the theorem in just 30 hours, with minimal human intervention. Fable 5.2 started from minimal prompts, built its own ways to coordinate, and used no mathematical references, only a list of intermediate theorem statements to score progress. It needed one-third of the action tokens as our first formalization and produced a proof seven times shorter. As can be seen in the figure below, Opus 5.5 (with the same 1,000-agent setup) completed only 26% of the formalization in 18 hours, and barely made any more progress after 30.
Most people don't really have intuitions on what it means to push the frontier now (because AI has surpassed everything they are good at, for example, coding). We are truly inside the singularity now (using Kurzweil's definition, slightly paraphrased - the time when technology progresses so fast no one can keep track with it.
This 1000x!
We've sort of entered this twilight zone where most humans don't have the time, expertise, or moral compass to be able to use these models.
We're all going to be harness engineers in the future.
I gotta admit, I am jelly as fuck I don't have access to a billion dollar datacenter. Would LOVE to put 1000 fable 5.2 agents to self-coordinate and get them to do something for me. Fuck life is unfair.
Does this coincide with the claimed nerfing of Opus 5.5? https://news.ycombinator.com/item?id=50422069
At this point, this is giving "She Deep Breaths More" from Onyxia in WoW.
Context, I had to look this up:
“…a reference to the repeated (and always erroneous) claims made after almost every patch that Onyxia was using her Deep Breath ability more often than before.”
https://warcraft.wiki.gg/wiki/She_Deep_Breaths_More_%2825_pl...
There were more than a few "leaks" saying it would be called Fable 5.5, so I'm curious if a) the leaks were just engagementbait or b) Anthropic changed it, despite the fact that Fable 5.2 being better than Opus 5.5 is confusing.
Surprisingly the model is available immediately in my Claude Code: usually there's a delay, either for rollout or safety.
> There were more than a few "leaks" saying it would be called Fable 5.5
Fed to someone specifically to test if they'll leak it and to confirm who that is?
I'm a bit out of the loop - would the engagement bait be becuase it is 3 minor jumps forward?
I assume because Opus 5.5 was a huge improvement over 5.0 (or was it 5.1? i forget). Fable still talks a fair bit like Opus 5.0 for me, so i'd quite enjoy a Fable equivalent to 5.5. Ie i'm quite interested in Fable 5.5.
Curious to see how 5.2 measures up in producing readable output.
Probably they have a more impressive model in preparation, and they're gonna dub that Fable 5.5.
Probably released on the same day they go public to generate maximum publicity.
Maybe it's just me, but I found Claude's writing style terrible compared to GPT or Gemini. It's "smart", but the way it communicates seems somehow off compared to other models.
Opus 5 and Fable 5 were unbearable in their writing style. They were very cryptic and hard to understand.
Opus 5.5 is a massive improvement.
opus 5.5 is still unbearable but better than 5.
I've found Claude's writing between 4.6 until 5.0 to be unacceptable if you are someone who is juggling 4-5 Agents. The Sol 5.X models all seemed to have great writing. Now with Opus 5.5 and Sol 6.1 I've found the trend is reversed -- Sol is now unbearable and Opus is workable again. Whatever caused the issues for Anthropic seem to be hitting OAI now.
I've heard a lot of chatter that agrees with you so anecdotally it's not just you.
Did you find that to be true for Opus 5.5 as well?
I couldn't use prior Claude models without rolling my eyes at the writing every few sentences, Astra was refreshing and then Opus 5.5 made even Astra look terrible in comparison.
I still find Opus 5.5 way too dense. MUCH better than Opus 5 which was unreadable.
Even the Plain English hack only is a temporary solution.
My own experience having used 6.1-Sol/6-Astra and Opus 5.5 extensively, Opus is more verbose and chatty. Multiple times I've gone "I'm not reading all that shit" when looking at Opus 5.5 output. This was also with the concise mode enabled in Claude Code.
Ever spoken to a savant programmer?
They're not typically known for their interpersonal skills.
Ethan Mollick's take and early testing with Fable 5.2: https://www.oneusefulthing.org/p/fable-52-and-jagged-agi
“Testers have found that Fable 5.2 and Mythos 5.2 communicate more clearly than past Fable and Mythos models. They generate outputs up to 30% faster than Claude Fable 5.1 and Claude Mythos 5.1, and they’re available at a lower price: $8 per million input tokens and $40 per million output tokens.”
Honestly incredible that they’ve managed to beat all benchmarks, while making it cheaper. Always happy to see improvements offered at lower costs.
Will be interesting to see what OpenAI does in response
They still need a response to Opus 5.5
It's really true.
I think frontier models feel terrible after a new model requires you to do less babysitting, it's just really hard to go back once you've had to handhold less.
The only way I use Astra and 6.1 Sol now is if they're being orchestrated by Opus 5.5 so I don't have to talk to them directly.
GPT-6.1-Sol is so incredibly sloppy and dumb, it's embarassing.
Context window maxes out at 200k in Claude Code instead of 1M on my $200 ultra plan? is this intended?
Is this the Model 4 internal name?
Reference: https://x.com/METR_Evals/status/2118412205662314935
Probably not, not enough time to red-team
I knew it, Claude was pretty dumb from last days until now
Looks like some plans get extra fable usage: https://support.claude.com/en/articles/17506494-claude-fable...
Assuming no implicit phacking a 15 point jump on tbench scientific is a big deal.
I suspect we will see people being adamant that this model simply 'feels' better due to long horizon goal alignment improvements ( like this kind that would cause this jump)
Gemini carbon low costs more than medium! Weird results on that chart, care to explain?
Those theoretical physics papers look like the real thing (i.e. not mathematical physics like the OpenAI Yang-Mills result). It looks like AI is finally branching out of math.
Its cheaper at $8 per million input tokens and $40 per million output tokens which is also pretty sweet.
Additionally, the terms for the free resets are unusually generous: https://support.claude.com/en/articles/17506494-claude-fable...
> the number of Fable resets you get depends on how long you’ve been a paid subscriber [...] Subscription length is measured from the date of your first paid bill. If you canceled and came back, the time since your first paid bill still counts.
> These resets expire on March 8, 2027 at 11:59 pm PT.
tl;dr if you joined when Opus 4.5 became a thing, you got 2 resets.
i might be wrong here but it seems these are not resets of your weekly quota but resets of the Fable limit.
That means you can lift your Fable max 50% usage and you'll just end up burning through your quota faster. You won't have more tokens.
This is what it says in the UI:
"Resets your Fable limit, and takes this week’s Fable use off your weekly limit."
"Weekly limit: gives you back the usage you’ve spent on Fable models"
It's pretty sad that they obviously have better models in the pipeline and don't release them yet for strategic reasons.
Too expensive for me
Really? A $200/mo plan gives you an enormous amount of inference on this model.
Doubling Remote Labor Index (RLI) - at 41.3% compared to around 20% for the other models - I wonder what kinds of new agentic work this is going to unlock.
I've very interested in the improvements to robotics control capabilities. This model was able to outperform specialized robotics models. It's the bitter lesson coming for robotics.
>> Fable 5.2 and Mythos 5.2 are a step above Opus 5.5 in almost all the capabilities we evaluated.
Was this before you lobotomized Opus 5.5 or after?
https://news.ycombinator.com/item?id=50419019
Definitely raises an interesting question. Are they using launch benchmarks for Opus 5.5, or Oopus 5.5 from this week?
Normally I would say "just check against the original Opus 5.5 benchmarks" : https://www.anthropic.com/claude-opus-5-5
...but most of the benchmarks don't line up anymore because the benchmarks have changed.
The Humanity's Last Exam result matches the initial announcement, though.
FrontierCode 2.0 is such a poor benchmark for Gemini since it's the only model forced to use Devin CLI while other models get to use their preferred harness.
I wonder why that team singles Gemini out like that.
Just read the other thread about opus being nerfed and thought "cue the new release"... paha
cool, and how good does it kill little girls in the middle east?
Why dont I feel any urge to even try Fable 5.2 and Mythos 5.2 unless it will force up on me?
Compared Opus 5.5 200 usd a month with Codex 200 usd (Astra High) and Kimi Pro 100 usd a month ... and now have a new conspiracy theory - more their suck your code the better outcome even just over few days time on large code base.
Horrible experience correcting course of Codex and Kimi flow compare to Opus. Opus much more responsive
Dont feel any gain of Codex 200 usd over Codex 100 usd
Kimi Pro feels like a fake Louis Vuitton bag with extreme high material quality and yet still fake ... a bit trash talk ... but clearly seen not so distant future where I can live without Codex and Claude
We'll see what my bank account has to say about this:
> Fable 5.2 at High effort scored about level with Opus 5.5 at Max effort, for roughly half the cost per task
I have gotten three (3!) banked resets. That is something.
I didn't realize that Opus 5.5 was more capable than Fable 5.1 in virtually every category.
Not anymore, they have nerfed it.
Very interesting that they set the flagship's version number to 5.2, but their penultimate model is 5.5. Do they want people to to continue to think that Opus is based on more advanced technology? The decision these AI companies make continues to baffle me.
google's agy usage is so generous, it might not matter
Leaps across the board but Robotics jump is huge.
god bless token resets!
> god bless token resets!
You have session limits?
You should probably give OpenCode a try. Can run DeepSeek, Qwen, Kimi, … but also Fable 5.2, Opus 5.5 and GPT 6 Astra
While I like OpenCode a lot, it’s not sensible to use it for anthropic models vs. a subscription (other than to use the included API credits) in terms of cost
The end of software is so close, very soon the code itself will be nearly obsolete, there will be schedulers, queues, a database, and the model
Which model though? Sounds pretty dystopian if there's no local option. Even speaking as a business, why would I want a environment where every single bit of information in my business goes to OpenAI or Anthropic?
It'll be cheaper than hiring people and maintaining software, presumably
because you wont be in business trying to compete by using human labor/open models
Why not? If we use labour as a comparison. The majority of businesses by definition do not hire the most incredibly intelligent and most competent workers and they are still competing in their niches and earning an income. Why wouldn't that be the case with choosing an open model vs fable 8.0?
+ 2 Fable resets HAHA. "Too expensive" who cares! Thanks Dario
stop OPUS 5.5 FUCK, DAMMIT, THEY DIDN’T JUST NERF HIM—THEY ABSOLUTELY OBLITERATED HIM, THE FUCKING MISANTHROPES! I EVEN MADE A MISTAKE IN MY ENGLISH IN THE PREVIOUS SENTENCE—NOW THAT FUCKING CLAUDE IS TEACHING ME ABOUT BUGS!
omg FINALLY lets GO