Claude Opus 4.7 • Hacker News Board

Re: Claude Opus 4.7

#161

Post by vunderba » Thu, Apr 16, 2026, 3:20 PM UTC

muzani wrote: post #29

buildbot wrote: post #12

Too late, personally after how bad 4.6 was the past week I was pushed to codex, which seems to mostly work at the same level from day to day. Just last night I was trying to get 4.6 to lookup how to do some simple tensor parallel work, and the agent used 0 web fetches and just hallucinated 17K very wrong tokens. Then the main agent decided to pretend to implement tp, and just copied the entire model to each node...

For me, making it high effort just fixed all the quality problems, and even cut down on token use somehow

This. They kind of snuck this into the release notes: switching the default effort level to Medium. High is significantly slower, but that’s somewhat mitigated by the fact that you don’t have to constantly act like a helicopter parent for it.

Re: Claude Opus 4.7

#162

Post by ttul » Thu, Apr 16, 2026, 3:20 PM UTC

TIPSIO wrote: post #9

Quick everyone to your side projects. We have ~3 days of un-nerfed agentic coding again.

... your side projects that will soon become your main source of income after you are laid off because corporate bosses have noticed that engineers are more productive...

Re: Claude Opus 4.7

#163

Post by erdaniels » Thu, Apr 16, 2026, 3:21 PM UTC

Kim_Bruning wrote: post #3

> "We are releasing Opus 4.7 with safeguards that automatically detect and block requests that indicate prohibited or high-risk cybersecurity uses. " This decision is potentially fatal. You need symmetric capability to research and prevent attacks in the first place. The opposite approach is 'merely' fraught. They're in a bit of a bind here.

Now we have to trick the models when you legitimately work in the security space.

Re: Claude Opus 4.7

#164

Post by stingraycharles » Thu, Apr 16, 2026, 3:21 PM UTC

Tiberium wrote: post #36

cupofjoakim wrote: post #20

> Opus 4.7 uses an updated tokenizer that improves how the model processes text. The tradeoff is that the same input can map to more tokens—roughly 1.0–1.35× depending on the content type. caveman[0] is becoming more relevant by the day. I already enjoy reading its output more than vanilla so suits me well. [0] https://github.com/JuliusBrussee/caveman/tree/main

I hope people realize that tools like caveman are mostly joke/prank projects - almost the entirety of the context spent is in file reads (for input) and reasoning (in output), you will barely save even 1% with such a tool, and might actually confuse the model more or have it reason for more tokens because it'll have to formulate its respone in the way that satisfies the requirements.

While the caveman stuff is obviously not serious, there is a lot of legit research in this area.

Which means yes, you can actually influence this quite a bit. Read the paper “Compressed Chain of Thought” for example, it shows it’s really easy to make significant reductions in reasoning tokens without affecting output quality.

There is not too much research into this (about 5 papers in total), but with that it’s possible to reduce output tokens by about 60%. Given that output is an incredibly significant part of the total costs, this is important.

https://arxiv.org/abs/2412.13171

Re: Claude Opus 4.7

#165

Post by _boffin_ » Thu, Apr 16, 2026, 3:22 PM UTC

aurareturn wrote: post #31

endymion-light wrote: post #16

I'm not sure how much I trust Anthropic recently. This coming right after a noticeable downgrade just makes me think Opus 4.7 is going to be the same Opus i was experiencing a few months ago rather than actual performance boost. Anthropic need to build back some trust and communicate throtelling/reasoning caps more clearly.

They don't have enough compute for all their customers. OpenAI bet on more compute early on which prompted people to say they're going to go bankrupt and collapse. But now it seems like it's a major strategic advantage. They're 2x'ing usage limits on Codex plans to steal CC customers and it seems to be working. It seems like 90% of Claude's recent problems are strictly lack of compute related.

You state your hypnosis quite confidently. Can you tell me how taking down authentication many times is related to GPU capacity?

Re: Claude Opus 4.7

#166

Post by Glemllksdf » Thu, Apr 16, 2026, 3:22 PM UTC

aurareturn wrote: post #31

endymion-light wrote: post #16

I'm not sure how much I trust Anthropic recently. This coming right after a noticeable downgrade just makes me think Opus 4.7 is going to be the same Opus i was experiencing a few months ago rather than actual performance boost. Anthropic need to build back some trust and communicate throtelling/reasoning caps more clearly.

They don't have enough compute for all their customers. OpenAI bet on more compute early on which prompted people to say they're going to go bankrupt and collapse. But now it seems like it's a major strategic advantage. They're 2x'ing usage limits on Codex plans to steal CC customers and it seems to be working. It seems like 90% of Claude's recent problems are strictly lack of compute related.

Its a hard game to play anyway.

Anthropics revenue is increasing very fast.

OpenAI though made crazy claims after all its responsible for the memory prices.

In parallel anthropic announced partnership with google and broadcom for gigawatts of TPU chips while also announcing their own 50 Billion invest in compute.

OpenAI always believed in compute though and i'm pretty sure plenty of people want to see what models 10x or 100x or 1000x can do.

Re: Claude Opus 4.7

#167

Post by Jeff_Brown » Thu, Apr 16, 2026, 3:22 PM UTC

skerit wrote: post #56

yanis_t wrote: post #25

> where previous models interpreted instructions loosely or skipped parts entirely, Opus 4.7 takes the instructions literally. Users should re-tune their prompts and harnesses accordingly. interesting

I like this in theory. I just hope it doesn't require you to be be as literal as if talking to a genie. But if it'll actually stick to the hard rules in the CLAUDE.md files, and if I don't have to add "DON'T DO ANYTHING, JUST ANSWER THE QUESTION" at the end of my prompt, I'll be glad.

It might be a bad idea to put that in all caps, because in the training data, angry conversations are less productive. (I do the same thing, just in lowercase.)

Re: Claude Opus 4.7

#168

Post by ls612 » Thu, Apr 16, 2026, 3:22 PM UTC

Kim_Bruning wrote: post #3

> "We are releasing Opus 4.7 with safeguards that automatically detect and block requests that indicate prohibited or high-risk cybersecurity uses. " This decision is potentially fatal. You need symmetric capability to research and prevent attacks in the first place. The opposite approach is 'merely' fraught. They're in a bit of a bind here.

Only software approved by Anthropic (and/or the USG) is allowed to be secure in this brave new era.

Re: Claude Opus 4.7

#169

Post by ACCount37 » Thu, Apr 16, 2026, 3:22 PM UTC

methodical wrote: post #122

ACCount37 wrote: post #80

> We are releasing Opus 4.7 with safeguards that automatically detect and block requests that indicate prohibited or high-risk cybersecurity uses. Fucking hell. Opus was my go-to for reverse engineering and cybersecurity uses, because, unlike OpenAI's ChatGPT, Anthropic's Opus didn't care about being asked to RE things or poke at vulns. It would, however, shit a brick and block requests every time something remotely…

To be fair, delineating between benevolent and malevolent pen-testing and cybersecurity purposes is practically impossible since the only difference is the user's intentions. I am entirely unsurprised (and would expect) that as models improve the amount to which widely available models will be prohibited from cybersecurity purposes will only increase. Not to say I see this as the right approach, in theory the two for…

Yes, and the previous approach Anthropic took was "allow anything that looks remotely benign". The only thing that would get a refusal would be a downright "write an exploit for me". Which is why I favored Anthropic's models.

It remains to be seen whether Anthropic's models are still usable now.

I know just how much of a clusterfuck their "CBRN filter" is, so I'm dreading the worst.

Re: Claude Opus 4.7

#170

Post by bensyverson » Thu, Apr 16, 2026, 3:22 PM UTC

Tiberium wrote: post #36

cupofjoakim wrote: post #20

> Opus 4.7 uses an updated tokenizer that improves how the model processes text. The tradeoff is that the same input can map to more tokens—roughly 1.0–1.35× depending on the content type. caveman[0] is becoming more relevant by the day. I already enjoy reading its output more than vanilla so suits me well. [0] https://github.com/JuliusBrussee/caveman/tree/main

I hope people realize that tools like caveman are mostly joke/prank projects - almost the entirety of the context spent is in file reads (for input) and reasoning (in output), you will barely save even 1% with such a tool, and might actually confuse the model more or have it reason for more tokens because it'll have to formulate its respone in the way that satisfies the requirements.

Exactly. The model is exquisitely sensitive to language. The idea that you would encourage it to think like a caveman to save a few tokens is hilarious but extremely counter-productive if you care about the quality of its reasoning.