https://www.reddit.com/r/ClaudeAI/comments/1sn57af/introducing_claude_opus_47_our_most_capable_opus/

Modified by: Mysticdrew on 2026-04-16 11:10:13 AM

I think they updated the leaderboards on tuesday.. maybe it will get added then

looks like tokan usage is way up, and large context has a major regression

@PitViper they reset claude 100% likely because of the issues wiht opus 4.7

I've read that 4.7 is both better and worse than 4.6. I havnt used it yet. Have you noticed an improvemnt?

Yes, it does seem better than 4.6, not a lot. But also uses more tokens i've noticed. They've been doing some udpates too, seems every day claude-code updates. I still have one session on 4.6, I hve not closed it yet.

it also refuses to do a lot more, and if you want to do any security stuff you have to apply for exemptions

Modified by: swiftraccoon on 2026-04-19 12:05:04 PM

was it a painful process to get the exemption?

nope. could've been worse. they want to know who you are, who you work for, and what you're working on and why.

oh, interesting. Is this for an actual company you work for or for your pantheon RE stuff?

i just put generic security use cases and put a not real company name (but do have it attached as a GitHub organization). my real company visible on my LinkedIn which they also requested tho. didn't require uploading ID, etc. also wanted my GitHub.

Modified by: swiftraccoon on 2026-04-19 12:12:52 PM

I wonder how closely they track the use of an account with a waiver.

Opensource giving ClosedAI a run for thier money.. If this pace keeps up.. Claude might have a real competitor!

I saw something, claude pro no longer has access to claude code??