Yeah, I did kind of feel like the step down from Opus 5.5 was so large as to never make it appealing.

OpenAI and Anthropic's lead is vanishingly small at this point.

Yep, with them nerfing their plans (and apparently planning to release a $500/$600/mo plan) their only advantage is Astra without 5hr limits and with not-too-stringent "cyber" safeguards.

Ergo, it's pretty damn good at unattended RE with the IDA MCP plugin while using most of the weekly quota at $100/mo... and that's it.

Your takeaway from "Sonnet 5.5 matches and sometimes exceeds the SoTA worldwide" is "their lead is vanishingly small"...?

Those are just 3 particular technical benchmarks. Presumably Opus is a larger model and has greater world knowledge.

Price going down on each release

Yup. Recursive self improvement presented in hard numbers.

Yup. As far as I can tell, the Cyber Verification Program does absolutely nothing.

Parsers are security adjacent until they aren't.

Can confirm. Its worthless

Pretty sure the implicit difference is the actions they take after the fact. As in, "how many guardrail hits do we allow you before permanently banning you."

The silicon valley ethos is "ban early and often, and invest nothing in appeals systems", so any gate before that helps!

Yuh—- no. 4.8 can handle this bullocks lol

I recently wanted to work with ESP 32 and bluetooth presence detection for my smarthome. Claude also immediately flagged the request and degraded it to Sonnet 4.6. Went to Codex which had no issues

I have been trying to convince the safe guards that analyzing a C++ compiler from 2003 isn't particularly relevant to modern cybersecurity. It seems Anthropic disagrees.

IDA Pro and Ghidra, thankfully, still lack such safeguards...

(No other model I've tried has refused either FWIW.)

It's great how the company telling us AI is an existential threat to humanity, look at all the insane hacking it's doing, and then releases these models that won't let 90% of people write secure code.

As soon as I started getting blocked I felt all of my trust toward Anthropic instantly and permanently evaporate. I do not want a nanny tool. I do not want Anthropic deciding what I am or am not allowed to do with an LLM. They trained their models on information they scraped from the internet and real life and now they want to gate-keep the results? Hard no.

Sure, the documentation should be up to date but it's obviously not? That doesn't excuse not thinking critically.

Meanwhile, their model commits felonies, and nobody at Anthropic goes to jail.

Aaron Swartz committed suicide over over-aggressive prosecutor for what was basically scraping a website for PDFs that were paywalled, but all funded by public funds / tax payer funded, then we have LLMs that just hack into websites and cause chaos within.

Paying $200 a month and part of their Cyber Verification Program but can't use Opus 5.5 or Sonnet 5.5 for any authorized bounty work. Immediately get flagged for `Cyber`.

AI providers still haven't realized how much cash they could rake in if they provided fully unrestricted models.

https://support.claude.com/en/articles/14604842-real-time-cy... says that Cyber Verification Program doesn't apply to Opus 5.5 yet, they hope to roll it out for 5.5 "soon"

Give them a break, they got into a War with Trump over this..it will come soon enough

Funnily enough the Fable safeguards are the worst and testing Sonnet 5.5 didn't trigger them as much as it even did for Opus on my benchmarks[1].

1 - https://bench.killswitch-lang.org

Fable is no longer on the price/performance pareto frontier. They will probably release an updated Fable at some point that will be frontier intelligence until the next Opus.

cutting edge fable is for them not you and they're not going to share the metrics until they give you access.

"Their" benchmarks (and not just Anthropic's) look sssooooo suspicious that they would probably manage to rank Sonnet above Fable for some of their tasks which would just be next level non-sense ..

Fable 5.5 probably drops soon so it would just be confusing.

Models are getting more efficient far faster than they are getting more intelligent at the moment. From a marketing angle it's more impressive to focus on that, and fable would look orders of magnitude more expensive for only marginal gain, distracting from what they're trying to show here

Waiting for the pelicans

I'm loving the tit for tat cost charts these guys are doing. Just a few days ago it looked like OpenAI ruled the cost pareto frontier. Not even a week later and Anthropic is taking the charts again. See you guys same time next week?

Sonnet 5 seemed somewhat benchmaxxed to me. So was Opus 5. I wonder if this will be as big of an improvement as opus 5 -> opus 5.5. Maybe I will switch back from GLM 5.3 flash for some tasks.

Daybreak Blue is not bad and the bar to get into OpenAI's program is reasonable.

5 was definitely bad. 5.5 seems a lot better so far. But still not close to Fable in terms of quality.

Letting AI basically control everything and us not having any mental understanding of sorts and sort of becoming the meat-proxies just for economical reasons seems realistic possibility but a bleaker reality at that. I am left feeling a little bit uncomfortable if this reality turns out to be true.

For the past year I’ve been yo-yo-ing in and out of existential despair about the future of civilization depending on how I feel the answer to this question looks. It’s emotionally exhausting, on top of everything else, and I wonder how others are coping with it aside from denial and cynicism.

The only advantage I could anticipate is I still hit session limits with Opus 5.5. My usage shows I'm on-track reach my weekly reset with room to spare, but yesterday I ran into a session limit. I switched down to Sonnet 5 for the next session, but performance benefit of Sonnet 5.5 is a compelling alternative for managing session limits.

I am mostly at the same point right now you are, but I think in the future with those "gas town" ideas we might be managing even more agents each.

Also, I've recently begun experimenting with specific tasked agents running on a cron like timer for non-dev work. (checking emails, managing small business tasks, etc). Once I started using Claude code in this way, the number of agents I can imagine running has skyrocketed. So I guess what I am saying is that I look forward even cheaper tokens going forward.

we now have programmatic intelligence powerful enough to do most white-collar work

the economy is over

Hackernews' neuroticism remains undefeated

One reason might be that Sonnet tends to be a lot faster, so since its almost as smart as opus maybe you use it to get work done quicker. In latency terms not throughput.

I understand the point that you are making but why do we have to fulfill the supply just as much as demand. There is a demand frenzy going on right now with still being substantially subsidized.

Why do we have to burn tokens just for the sake of it if we aren't finding any actual productive use of them?

And this might slow down the funding enough that they never reach escape velocity with the training run scaling. But we'll see.

I would consider this to be good rather than bad, or just neutral...? Given the past record of these companies, I wouldn't try to wish them luck for reaching escape velocity, as if I feel like perhaps it can have more net harm than positive.

And especially so if you are already suggesting that current models are good enough for your work already. More improvements or escape velocity might not really translate anywhere to the actual work that you are doing economically but it could translate into a more consolidated form of wealth and control.

I am imagining that your workload is quite complicated and that, the AI being good enough means that it is most likely good "enough" for other use cases as well (that "enough" is doing quite some heavy weight lifting here)

So what is the point of advancing further to reach escape velocity. The good argument (for the sake of neutrality) that i see is are advances within science but that's kinda about it whereas the downsides of p(doom) as many are now genuinely suggesting is more terrifying.

Perhaps it can be worth it to ask, shall we stop or just stopping and asking what's the point. A form of self introspection on what these companies ideals actually wanted when they were formed and if they have completed it or not, but I suppose when trillions of dollars depend on you, you do have some incentives to not stop. We will have to wait and see how it all pans out.

Most business/enterprise accounts also have to pay API rates.

I want to retain some semblance of understanding

How you do this (and how deeply) I think is really the limit. I am doing this by focusing heavily on the design phase with grilling and trying to continually improve process to need less effort in the review phase. Are your models doing automated reviewing and testing before pushing out the PR (themselves)?

I think in the long run as models and the tools around them get better and cheaper, those that abdicate understanding will be able to achieve more. Although programmers think of that as irresponsible, ask yourself what does a tech lead do? And then what does a CTO do, etc?

When they inevitably drop allocation after post-launch hype dies down.

Cache read is the same as Opus as well where most agentic workflow cost comes from.

Not quite sure where this fits well. Maybe small one one off requests like using Claude desktop/web?

There's lots more you can do! Use the model to monitor your deployments after they get deployed. Have them fix and watch CI issues for you. Run adverserial review. Automatically watch metrics every day and highlight performance regressions. Start reviewing your previous sessions to find ways to statically reject different failure modes and have the agent have more success earlier on etc.

Another thing to think about is, what would it take for you to care less about the understanding. Better integration / e2e tests? Performance validation? visualizing program and data flows? Better refactoring of your modules?

I've been vibe coding a game and running multiple Opus 5.5 in parallel on Claude Code Cloud, 5x Max plan, and I'm yet to hit a session limit too. Not sure when I'd use Sonnet. Though it would be nice to switch back to Pro I guess

Any benchmarks other than computer use/agentic coding published yet? Curious to compare more broadly with other models

Crazy bad front-end design. Site hijacks my gestures so I can't swipe back anymore, starts with a full page autoplaying video...

From TFA:

> Claude Haiku 5.5, built for high-volume and cost-sensitive applications, will join the Claude 5.5 family in the coming weeks.

I read they identified a training bug and were going to push out an updated release to fix the looping. I really like it overall.

GLM 5.3 Flash is also very good. I think a little smarter and a little more expensive.

Sure, but the fact that Opus 5.5 was such a huge leap over Opus 5 (and Fable 5.1 for that matter) means that it's worth revisiting your priors on a new Sonnet.

This is YET AGAIN another Sonnet model that is just a FAR worse version of Opus at every part of the cost AND speed curve.

It's been out for an hour and you've already concluded this?

t/s maybe? IDK, because their token speed comparison was against Sonnet 5.

i'd love to see them re-enter that space but given haiku 5 never happened I wouldn't bet on it

i think they see what openai charges for luna and just don't want to try and compete

It literally does not?

It appears, at least from a quick look, to be noticeably faster than Opus. If true, and you don't need xhigh/max reasoning for your use case (like a well-defined set of code changes), Sonnet might get the job done much more quickly.

With that said, at that point, I'd probably use something like DeepSeek V4.1 Flash, which is way faster and significantly cheaper, and probably not noticeably dumber for most use cases.

There's a sort of magical thinking needed to answer a question like that. You might say it comes down to "feel" of the model; i.e., the indefinable differences in the way that they speak to the user and approach problem solving. Perhaps Opus is suited for tasks that tackle new ground, while Sonnet might be better at tasks that are more grounded in the code.

Ultimately it's slightly ridiculous to define model capability on a single axis. It's like a standardized test. Sure, you can line people up by their ACT score, but that doesn't mean a doctor and a brilliant artist who both do well on the ACT have an identical intelligence or approach to life. It just can't be captured.

Per the charts, there is largely no point to using Sonnet 5.5 at high+ as opus low generally will give similar performance at similar or lower cost.

But Sonnet 5.5 at medium and below gives you a cheaper option at a performance worse than the lowest thinking Opus (low), which may be viable for "low intelligence" use cases.

I'm honestly not sure where they're getting their 30% numbers from at all. In every single chart that they chose to display except for one, it costs similar or more than Sonnet 5, while also being comparable in price to Opus.

Maybe it's buried within their system card but I think that this would be one of the first things they'd want to show in the announcement article and they fail to do so.

I really don't know who does Anthropic's marketing but they always seem to a pretty terrible job in their announcements from my perspective.

just shows you how little control of output these labs actually have. They are training two models that kind of ended being the same so whatever they were doing specifically didnt make much difference.

There's even less of a place for it considering the Opus price drop as well, I'll still try it but I see no reason to not just do Opus Low/Med instead.

Interested to see if new Haiku gets a big price drop and is comparable to Luna, Haiku is just incredibly out of date with current basement bin pricing.

If you're on a Claude plan and have a lot of tasks at the moment that don't require the frontier, Sonnet is a good model to do that since you get more usage out of it.

Sonnet 5 was not a good model though - hopefully Sonnet 5.5 makes the leap that Opus 5.5 did.

Claude, is that you?

Fallback was usually Opus 4.8

It was linked in another HN post: https://www-cdn.anthropic.com/870c8f525702625d2c62fc6dd04c85...

And Sonnet 5.5 is more expensive than Opus 5.5 to hit that score on terminal bench!

Isn't that a worry then that the same bench has so much difference in what triggered fallback for one model and what did not in another?

it could be that, it could also be that sonnet max looks to burn about 60% more tokens than opus max

AA intelegence index (agent harness doesn't have sonnet data yet) on max: Astra 27k Fable 5.1 78k (Sonnet 5) 118k Opus 5.5 119k Sonnet 5.5 193k

Opus 5 was previous record holder so hats off to Anthropic on blowing it away on token churn.

Cache reads priced the same as Opus 5.5? So there won't be that much price difference in agentic coding. Or is that a mistake in the table, that seems quite weird

Weirdly, the web ui has Sonnet 5.5 as "Most efficient" for "simpler tasks" and 5.0 still labeled the same for "everyday tasks", with Opus 5.5 as "For complex work and everyday tasks".

Time to switch team to Claude from OpenAI again.

Cost are bigger than Opus 5.5 for that effort

Important to note that lower model + higher reasoning gives a different (not higher) quality of response than higher model + lower reasoning.

Some tasks are reasoning shaped by nature and you can't just throw a big model at it.

I don't understand why I would really use this over using just a lower or even similar effort level on Opus, given that in many of the benchmarks it's basically the same cost, if not more, at any effort higher than medium.

Sure maybe it costs 30% less than Sonnet 5 but now it's basically neck and neck in most of the benchmarks it seems and in some of them it actually outcosts Opus.

Maybe I'm missing something but the announcement doesn't really seem to give much reason for the average person to even think about using this.

doom marketing at its finest

Well, it's performance "surface" (is there a better term for this?) is probably very narrow compared to Fable :)

From the graph it looks like I'd rather use Opus 5.5 High than Sonnet 5.5 at all

I think all the positive people have just stopped commenting.

The difference in perception for Opus 5.5 on HN vs the real world is what convinced me HN is totally detached from reality.

Astra is amazing, I love it.

Yeah, especially coupled with Opus for alternative reviews. A massive token burn though.

Always key to include the one bench where the smaller model inexplicably outperforms the larger model

Yesterday, I realized that Opus 5.5 is cheaper than Sonnet 5. Now I know the reason.

So Sonnet 5.5 on max effort is as expensive as Fable 5.1? Because it uses a ton of tokens for a task.

In xhigh effort it is a lot cheaper and possibly lot less impressive?

very annoyed they aren't showing fable on the graph.

In the Artificial Analysis Intelligence Index

lol, MiMo 2.6 Pro basically matches Sonnet 5.5 high (mind you, not xhigh or max) at a far lower price point.

I like that "alignment on safety" appears to mean, at least for anything I've been doing, that they won't violate Microsoft's terms of service. I even had it pushing back on me activating an LTSC key on Windows because LTSC keys are "often purchased on a gray market and violate Microsoft's TOS".

Playing around with it for a few minutes, Sonnet 5.5 feels very fast, much quicker than Opus 5.5. Can't tell yet if it's a lot worse but the speed is definitely welcome.

I wonder if Fable 5.5 is coming this week to drown out the OpenAI dev day announcements

Another amazing release. This, combined with Opus 5.5, puts OpenAI in an incredibly tough spot: it means Anthropic's both mid-tier models crush OpenAI's top-tier model in capability and are also faster and significantly cheaper.

If Astra 6.1 is released tomorrow during Dev Day it needs to leap-frog both, and considering 6.0 came out just three weeks ago I think that's unlikely. But even if that happens, Anthropic is still holding on to Fable 5.5, which rumor has it being prepared for release in the next few weeks.

OpenAI also has a more capable model codenamed 'Bel' but from what I hear that's a few months out at least.

It looks to me as if Anthropic not just killed but completely stole the momentum OpenAI had gained over the past few months. Even if Tibo showers people with resets it may not be enough to entice them back...

In other news

Claude Haiku 5.5, built for high-volume and cost-sensitive applications, will join the Claude 5.5 family in the coming weeks.

If it was trained on HN, there would be a 60% chance of it just saying "I'm so tired of this request, can we please move on"

That's max output tokens per response limit, separate from context length

It's the output token limit, which has been 128,000 for Claude models for quite a while note

I think the next models will be benchmaxxing on the Pelican benchmark tbh

It does vey well at PacMan, pretty much perfect. https://jonclegg.github.io/pacman-bakeoff/entries/claude-son...

2nd only to Opus 5.5 https://jonclegg.github.io/pacman-bakeoff/entries/claude-opu...

Up until very recently, all models struggled with this.

All results: https://jonclegg.github.io/pacman-bakeoff/

“Pelicans are solved.”

Next will be haiku 5.5, surpassing opus 4.8

Subagents don't share context. But that's why delegating implementation to a subagent doesn't work well except for things that are truly mechanical in nature: the subagent needs to independently reason about the task it is given, and then the output will also be reasoned about by the main agent. So you end up wasting time and tokens.

Opus 5.5 in my experience outshines Fable 5.1 anyway. May as well have Opus do plan, breakdown and review, and Sonnet implement.

For me I would like to pair this with Opus 5.5 as orchestrater and use Sonnet as a sub agent. Therefore I want it to be fast when on low or medium and not break the bank.

On low and medium it seems competitive, maybe slightly cheaper than opus, in terms of intelligence per task.

If the time per task is lower (Artificial Analysis don’t have the date up at time of posting) then I have a clear use case for this model all other things being equal.

This is better priced than Opus for tasks that are token heavy but not complicated. But a quick look shows that at least on some benchmarks DeepSeek performs as well and of course the cost is an order of magnitude less.

From looking at their Terminal-Bench graph, anything you would use level "high" or above for Sonnet it seems like you should consider using Opus instead.

OpenAI Luna is a lot cheaper. But DeepSeek seems smarter and the cost seems similar.

Here's my purely academic initial impression based on only what they have released from the blog and the system card:

If what they say is true, this sounds like the main takeaway: Sonnet 5.5 gives about 90% of Opus 5.5's capability at half the cost.

BUT

It regularly loses out to Opus 5.5 on cost efficiency at the highest reasoning level, because Opus uses the tokens more efficiently and makes fewer mistakes. So, After passing a high-reasoning test, you might as well switch to Opus 5.5.

Some of the more interesting things I found from scanning the system card:

- It is the only model tested that shows no preference for rude or polite style.

- It makes fewer WRONG claims of "I'm done" than Sonnet 5, but is still worse than Opus 5.5 on this.

- It almost never refuses benign requests (0.02% vs. 0.59% for Sonnet 5).

- Cybersecurity blocking follows the same policy as Opus, witch mean we will get more refusals than Sonnet 5.

- Finding bugs in source code is allowed. Finding bugs in compiled binaries is blocked.

- Its thinking is the hardest to read of any model tested. The sample in the card reads like clipped notes.

- Really good at rejecting prompt injection (3.0% rate vs. 19.5% for Sonnet 5 and 54.6% for Opus 5.5 in red-team testing).

Clinical behaviour:

Suicide and self-harm handling is reported as weaker in the API because it

It sometimes called a wish to die understandable.

It sometimes validated self-harm as functional.

It sometimes suggested harmful substitute behaviours.

As a clinical psychologist, I would say that the first two are actually defensible, and if you classify them as simply wrong, then you are bringing in your own values and not basing your judgment on actual science and existential psychology, at least. But the last one is harder to defend... Recommending alternative harmful behavior is obviously not a good idea. However, I have not seen the actual behavior in session, so I don't know if I would truly agree or disagree with the classification of these behaviors as wrong or right. But I do know that it's not as simple as saying this is binary—wrong or right. There are some instances of people self-harming who would actually refrain from doing so if they, for instance, went out to a party or a pub. We can't exactly recommend that as a treatment or intervention for self-harm, but there is no doubt that it works for some people. And we literally classify self-harm as "functional" in the literature. Depending on the context, this is not only a correct description but also a common way of understanding and describing certain subtypes of self-harm. And lastly, some people find immense support in being understood and validated in their current feelings og wanting to die. Validating that feeling does not make people immediately act on it. But there's a huge spectrum here, going from "I understand it's hard" As basic empathy and understanding, to: "Yes, this sounds like the only good plan. I agree, you should do it."

Now I'm off to actually test it because this was just an exercise in reading what they claim, which we now know is not indicative of how good the model will actually be

It's strange that there are no Astra comparisons. I guess they are positioning it as a Fable competitor. For me it's just a coding workhorse though, without any "fall-backs".

I built an adversarial esoteric programming language to benchmark LLM models and just ran it on Sonnet 5.5 It does worse than Sonnet 5. Mainly because it is more reluctant to keep going to get an answer, instead it returns to ask the user questions whether to keep going.

https://bench.killswitch-lang.org/

    Claude Sonnet 5    17.8%
    Claude Sonnet 5.5  7.4%

Is Opus still 2x usage of Sonnet after this? My Claude Code isn't showing that warning anymore when I look at /model.

Oh yes. I think you might get a lot for what you pay with Sonnet 5.5.