Rage Has anyone tried chatgpt 5.6?

[metacognitivemaxx]

Apparently the sol or whatever, set to max effort
Is on par with fable and outperforms newly released grok 4.5 (grok is irrelevant here), and glm 5.2?

Has anyone tried it? It got released a few days ago finally after the us government and top US based companies had access to it for weeks.

Also what do you think of Gemini 3.5 pro, does it stand a chance? Will it top the charts? It's been fucking delayed for 2 months now I think

Or is google deepmind out of the ai race?

Please tell me there's atleast one high iq individual into llms here, I want to meet him to discuss a project I want to work with involving llms...

 

[metacognitivemaxx]

Current reasoning benchmarks

 

[rand anon]

I dont use Gemini or grok, but I don’t think it’s as good as fable yet

 

[metacognitivemaxx]

I dont use Gemini or grok, but I don’t think it’s as good as fable yet

Its fucked honestly, the us government and elite corporatocratic society had access to mythos for weeks before releasing fable:a subversion of mythos with "security guardrails" (got jailbroken in 2 days at initial release) to the public.

A new top ai model that is a "worldwide cybersecurity risk" doesn't worry me. It's the fact that only a certain group of people would have access to it, and are able to regulate it.

Btw anthropic will introduce mandatory id verification within a few weeks you reminded me of how much its fucked. Open source ai must win to until 2030 or the next 200 years are cooked