https://www.bleepingcomputer.com/news/artificial-intelligence/anthropic-web-config-hints-at-claude-sonnet-4-and-opus-4/
"Anthropic is secretly working on new models called Claude Sonnet 4 and Opus 4, which are believed to be the company's most advanced AI models.

This is according to Anthropic web configuration files, which were recently updated with references to Claude Sonnet 4 and Opus 4."

Very strong sign of claude 4 soon.

Would Claude 3.8 count as Claude 4?

@AugustSha no

This market is crazy. Let's do simple base rate calculation.

How many times has Claude 4 come out? Zero.

How many weeks have there been? 13.8 billion years x 52 = 717.6 billion weeks

So base rate for Claude 4 coming out on a given week is 0 / 717.6 billion < 1%.

"Stubbornly, I will believe, I will believe (Keep dreaming)"

seems to me like releasing Claude 4 tomorrow would actually undermine their developer conference. I don't think they would try to align them in timing.

@JoshYou this is fairly convincing that a new Claude will come this week, but updates me a bit towards the name being 3.8, if Neptune is a reference to 8.

all of the evidence I've seen so far has felt iffy and highly speculative - none of it from people involved (although Anthropic seems to keep its employees quiet about alluding to new stuff moreso than Google DeepMind). I don't really think that they're going to release a new model, just a new safety paradigm

via @SanghyeonSeo - could be something

https://x.com/btibor91/status/1925084250107478506

it’s definitely something

@Bayesian place limit orders then

@jim But that’s work. Am i missing something?

@Bayesian i dont think you're missing anything, but seems like ppl want your action so...

no, but they'll release another banger blog post on AI safety

https://www.anthropic.com/events

May 22, 2025 — 9:30am PST

Code with Claude 2025

Join us for the livestream of the Code with Claude opening keynote.

i give 1000m