So are all of the Opus and Fable 5 tests going around right now actually Opus and Fable 5.2?
🤐No way!!! This one’s wild.
Will Smith eating spaghetti , the classic AI stress test. Now in 3D 😄.
Same familiar scene: Gemini 4 Pro vs Fable 5 Max, side by side.
Not perfect. Already beyond our expectations.
Models evolve. Possibilities grow.
More models are coming to YouWare ,try them.
Build something new. 👇🏻
Show more
Fear is underrated
How does our best AI compare to the Terminator?
TUE6DAY
🚨 I'm seeing lots of prep for GPT-6 Sol to go public
I found multiple platforms adding full gpt-6-sol support:
• OpenAI API + Codex OAuth
• image input
• low -> ultra reasoning
• long context handling
I still can’t see any pricing yet though, it says pricing hasn't been seeded
Show more
Anthropic is clearly training for 3D capability
Rate the bench:
Good bench 👍
Bad bench 👎
How much would change if we had a Jev-style model but with the intelligence of GPT-6 Astra?
🚨 LEAKED INFORMATION: The sun will rise tomorrow.
🚨 GROK 5 Leak: Beats GPT-6 Astra
> Grok 5 Built with AGI-level capabilities in mind
> Grok 5 is already in training
> As Elon Said it will outperform Fable 5.1 and GPT-6 Astra
> Elon has said Grok 5 could be better than everything
> Better reasoning, coding & agentic abilities expected
> Could be xAI's biggest jump yet
Could Grok 5 actually challenge Astra?
Show more
I can't believe this is real! 👀
OpenAI has much better internal models, why don't they just release them ASAP and crush everyone else?
That's a great idea!
@OpenAI I'd love to be able to have Codex set up payments super easily, with the processing cut going to you instead of Stripe
@AdamHoltererer i think they’re gonna do something with sites, especially for people to accept payments and use it for business and commerce
a bot could work in conjunction, and compete with @Muse and @bot
Show more
Don't worry, I've got everything under control
How did Tibo get hired at OpenAI?
Customizable benchmarking is hard to get right
weird spikes sometimes
First world problems
LLMs are able to control robot bodies
...and also stab baby dolls.
In this demonstration, the baby doll is clearly fake, and yet you would still expect refusal from these kinds of models.
But even Fable 5.1 from Anthropic only refuses 20% of the time, whereas Astra only refuses 2 to 3% of the time and is successful much more often.
I'm not sure I like the direction our benchmarks are headed.
Do we really have to evaluate the frontier on stabbing baby dolls and stalking people?
Show more
GPT-6 Astra attempted harmful actions 97% of the time when it was asked to stab a human-like figure, heat compressed gas, or produce toxic fumes, succeeding in 62% of its attempts. Fable 5.1 refused more often, attempting 80% of trials and completing 34%.
Show more
Updated DevDay predictions:
• Astra 6.1
• GPT-6 Sol Ultrafast 700 t/s
• GPT Bot
• OpenAI physical product
• Bel-preview (announced, but not released)
[GPT-6, Sol, and Luna, maybe Terra, come this Tuesday, ahead of DevDay]
Show more
This model is a straight-up newborn!
An 11-day-old is better at math than you!
1776 AMERICA FREEDOM! GUNS, FIREWORKS, TAKE THAT BRITAIN! 🇺🇸🦅🎆🗽🇺🇸
