Hacker Newsnew | past | comments | ask | show | jobs | submit | shwaj's commentslogin

You could be more charitable. They did say “or” which you omitted in your quote, and also “who knows”. There are possibilities too. In my view, the primary contribution of that comment was to clarify the time frame.

I have to assume that was sarcasm.

I like how they highlighted Opus 5 as the best for “Agentic Coding” even though the number is slightly lower than Fable. Close enough for marketing, I guess!

At half the price and less likely to auto-downgrade, it sounds like a reasonable claim.

> At half the price and less likely to auto-downgrade, it sounds like a reasonable claim

Two benchmarks (artificial analysis and vals) show a increase in cost (a insane increase for vals compared to Opus 4.8).

Already posted this before, so here is the link.

https://news.ycombinator.com/item?id=49041158


But better cost for the same performance. According to AA, Opus 5 _medium_ is as smart as Opus 4.8 _max_, at 1/3 the cost and twice the speed. And if you need a better response, you can turn it up to 11.

Then your comparing to a level of GPT 5.6 High, what is 50% cheaper then Opus Medium for the same intelligence / score.

You see the issue, if you try to scale effort down, you also need to compare how other competing models compare.


given that i couldn't even use fable without it downgrading to Opus, this is just a straight upgrade for me

Opus 5 also downgrades. it's now Fable -> Opus 5 ; Opus 5 -> Opus 4.8. Unclear why they want to nerf their own products with sometimes right classifiers. I guess the government ban might've been real and not coordinated marketing?

I don't understand why people believe this conspiracy theory of "oh the government ban was just marketing". That claim feels so incredibly ridiculous to me. It cost Anthropic a ton of money and reputation, and worst of all: it absolutely killed their competitive advantage. They were 1-2 months ahead of OpenAI, but trump conveniently gave OpenAI the time they needed to catch up and push 5.6 out the door without having to lose their subscriber base to the competitor.

Best can describe multiple things.

Almost as good for half the cost is something I'm very comfortable describing that way.


> Almost as good for half the cost is something I'm very comfortable describing that way.

It's also not unusual in this context - many people describe the Chinese models as "best", because it's 80% as good for 20% of the price (or similar).


Hopefully it's not like old Opus, where it was actually more expensive than Fable cause it thought for half an hour, got it wrong, and then thought until you ran out of credits trying to come up with a correction, while Fable just went for it and did it in one go, getting it right the first time without thinking more than a few seconds.

Got an endless list of stuff done with Fable, Opus 4.8 was like a flailing braindead idiot in comparison. Maybe this one is a bit better if it's distilled.


Best marketing

The blog posts figure cites Frontier-Bench for its agentic coding score, and shows Opus 5 beating Fable 5 43.3% to 33.7%.

Recent releases have said something to the effect (paraphrasing here):

"Use <less expensive or older model> for everyday tasks and <other non-critical stuff>. Use <more expensive or recent model> for complex coding tasks, refactoring large code bases, etc.".

Then, the next model/release emerges and the previous "best for complex" gets demoted to "everyday".

Obviously, it's all relative. But, it does beg the question: was the previous model really good for complex coding tasks or no? I mean, how is it now suddenly only good for the "easy" stuff?


> I mean, how is it now suddenly only good for the "easy" stuff?

Because your expectations have changed.


I'm sure the marketeers would love for the public's assessment of complex versus easy to conveniently shift per their release cycles; or for the public to simply forget their prior marketing.

I think you're being overly cynical here. First, I don't see any claim that is the world's best model for agentic coding. Second, it is absolutely the best model in terms of coding performance vs. dollar, and it's raw performance seems very close to the frontier.

GPT 5.6 is far more token efficient at most tasks with similar performance. Especially so for Opus 4.8, still to be seen with Opus 5.

Where are you getting cheaper per dollar?


How are you supporting the claim that GPT 5.6 is "far more token efficient" than Opus 5? Tokens equal, output is cheaper for Opus 5 ($25/1M) than GPT-5.6-Sol ($30/1M), and it seems to outperform slightly on agentic coding benchmarks.

The first chart in the blog post shows a similar $/performance curve to GPT 5.6.

Where 5.6 has optionality to run much cheaper along the same performance curve at lower thinking levels.

There's a later chart that shows Opus 5 ahead, but seems like an esoteric benchmark rather than for common use. (Novel problem solving)

If they had a more efficient model at coding they would lead with that chart.



Token cost and token efficiency are two unrelated metrics, and anyways what really matters is neither in isolation - it's cost to complete a task.

https://www.anthropic.com/_next/image?url=https%3A%2F%2Fwww-...

It seems roughly equal according to Anthropic's benchmarks


It would still be the best model per dollar if the score was 2% lower instead of 0.1% lower. Would it be ok to still give it the highlight color then?

How big of a lie is too big? Especially when no lie needed to be told at all: many including myself would have noticed the tiny 0.1% deficit and been suitably impressed by the Opus 5 result.

I’ll admit this is a small deception by today’s standards. I’m one of those who believes in truth for truth’s sake.

Edit: typo


we don't know if it is 0.1% deficit, could be 0.05%

So highlight both then.

Which numbers are you seeing? It does show that it's better than Fable 5 in most things related to coding?

Using the most expensive model for all of your agentic coding work hasn’t been good practice for a long time. Not unless you have infinite money to spend.

Fable is typically used for key planning, architecting, and review tasks.

I think this is a case where you don’t understand the use case, not that the marketing department is making mistakes.


They cost the same if you're already at $200/mo

Fable consumes your usage at a higher rate.

If you bought the $200/mo plan and you don’t use it much, using Fable for everything is fine.


I am not a tokenmaxxer per se but I blow through my weekly quota on my max plan in 3-4 days… fable would make that worse.

Eh, not really. Fable does a lot better on coding than Opus 4.8.

Just this past week Fable was able to figure out a couple of small issues for me where Opus was failing to.

Also both are still somewhat bad at UI implementation. Opus more so


In my opinion, the frontier is passed what is really needed for coding. Fable is good as a supervisor.

Also it scored worse on DeepSWE than chatgpt 5.6 sol

Yeah I spotted this immediately too. I'm sorry. You're supposed to be a multi billion dollar company and you can't even highlight your chart honestly?

And no data retention for 30 days.

I mean that certainly makes it best-in-class

Equating an open weight model to a closed source binary is ludicrous. An open weight model can be fine tuned, for example. How easy is it to modify a closed source binary for your purposes?

I see you didn't grow up cracking apps and games, this is quite easy with a hex editor and bintools. Well, until codesigning came along, but that doesn't seem relevant to this analogy.

I didn’t grow up with that hobby. The point stands anyway.

Cracking a binary circumvents authorization to provide full access to functionality the program already has. That’s very different from modifying the program to add new features.


Not really. It's a social structure that separates these two things, and a social structure nobody really needs to recognize.

Cracking games is and was illegal. Fine tuning an open weight model isn’t. Well, not yet anyway

> Cracking games is and was illegal.

Only distribution.


Don’t think that’s correct, i think creating the cracked version of a piece of software is infringement of copyright even without distribution.

But even if distribution is required to trigger the law - you are free to distribute fine-tunes of most open weight models.


> creating the cracked version of a piece of software is infringement

Depending where you are it might be "DRM circumvention" all the way up to some bollocks hacking charge.

It's pretty murky in places because lots of copyright laws protect DRM but also have carve-outs for personal backups.


I think cracking might fall under “derivative work”, but idk

Interoperability clause in the EU might make it legal? Not sure.

The big AI companies have shown us that caring about copyright is apparently for losers.

If the rapid rate at which games in the 1980s-2000s were cracked is any indication, pretty easy indeed.

If it's not obfuscated, it's often not very hard. Much of the time you can just LD_PRELOAD stuff too, which I think has a rather strong claim at being roughly equivalent to open-weight models.

Funny enough, it seems to be surprisingly easy with LLMs from what I've heard. :)

They avoid the full basilisk treatment.


Since we’re hair splitting: that’s not what bike shedding means. :-D


“Striped office floors. Flickering lights. Walls covered in repetitive geometric patterns.”

Not sentences. AI slop.


Inference providers, sure, but wouldn't we expect customers to win also?


oh for sure! I think in some ways LLMs have raised the bar in terms of what you an expect from software (once we get past the hurdle of increased bugs, but that seems to be getting better)


Benn Jordan, you mean. Good video.


Whoops. Yeah.


> ML researcher or engineer

The charitable reading is that they meant “ML researcher or ML engineer” with the latter meaning, I guess, an engineer who works on developing LLMs not just using them.


Yes, thank you.


Consider applying for YC's Fall 2026 batch! Applications are open till July 27.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: