Sabitlenmiş Tweet
I benchmarked GPT-4.5 (ChatGPT) on the MineBench prompts before it's too late. Here are the results.
English
Raphi-2Code
17.6K posts
@Gc_qube have you tested whether its 5.6 pro
ask it about its juice
English
🚨Sonnet 5 is worse than Fable
🚨GPT 5.6 pro is worse than Fable
🚨Gemini 3.5 Pro is worse than Fable
And those models haven't even been released yet😭
x.com/synthwavedd/st…
leo 🐾@synthwavedd
🚨 SCOOP(s): - GPT-5.6 has been delayed and will no longer release this week. New target is ~mid-July. - DeepMind are not satisfied with the current state of 3.5 Pro and it will no longer launch this month. - Preparations for the launch of Bidi, OpenAI's new voice model, are underway in ChatGPT and we could see it available as soon as this week. - Claude Sonnet 5 is currently available for select enterprise customers under an Early Access Program and is seen as a stop-gap as progress on getting Mythos/Fable 5 back out have stalled. A bit of a disappointing end to the month, but July should prove more fruitful!
English
Raphi-2Code retweetledi
GREAT NEWS 🚨: Bidi 1 can now count from 1 to 100 without making up excuses and refusing.
This was a big issue with the previous voice model, so I am very excited that this is solved.
Huzzah!
English
@JaydenDavisNC I’m wondering why xai is going backwards
Raphi-2Code@R2Cdev_
What happened since Grok 4? 💀
English
Grok Build v0.1 has scored a
36% on Humanity’s Last Exam.
This is on par with the recent Chinese open weight models that are available to use.
xAI right now is a year behind in commercially available models.
What the heck happened?
English
🚨 SCOOP(s):
- GPT-5.6 has been delayed and will no longer release this week. New target is ~mid-July.
- DeepMind are not satisfied with the current state of 3.5 Pro and it will no longer launch this month.
- Preparations for the launch of Bidi, OpenAI's new voice model, are underway in ChatGPT and we could see it available as soon as this week.
- Claude Sonnet 5 is currently available for select enterprise customers under an Early Access Program and is seen as a stop-gap as progress on getting Mythos/Fable 5 back out have stalled.
A bit of a disappointing end to the month, but July should prove more fruitful!
English
IT IS LITERALLY STILL WORSE THAN THE 4O DEMOS FROM 2 FUCKING YEARS AGO WHAT ARE THEY DOING
🚨 AI News | TestingCatalog@testingcatalog
Jokes, tone change, and roaring 👀
English
I benchmarked GPT-4.5 (ChatGPT) on the MineBench prompts before it's too late. Here are the results.
English
I created a prompt to max out a model’s creativity.
Here’s the result for GPT-5.6 Pro (Prompt in the comment section below):
English
@RijnHartman Yeah, they said that actually, but the plan didn’t work apparently.
Maybe because of Fable 5.
English
And also probably a smaller improvement compared to the improvement between gpt-5.4 and gpt-5.5.
AiBattle@AiBattle_
GPT-5.6 has been delayed after originally being planned for release this Thursday The wait since GPT-5.5 is the longest stretch without a major GPT-5-family release since the 97 day gap from GPT-5 to GPT-5.1
English
Grok 4.4 has finished training (pre-training) - release ETA is mid-June 2026 - and moves Grok from 0.5T to 1.5T, a "major upgrade" with the added Cursor data to be added soon.
Are we finally getting a SOTA xAI model?
Elon Musk@elonmusk
We are improving the 0.5T Grok foundation model V8 (public version 4.3) every few days. The 1.5T V9 just finished training (incorrectly called pre-training) and is a major upgrade. Next, we are adding the Cursor data in supplemental training (others call this mid-training), then SFT and RL. About 3 or 4 weeks to release. This will be a banger.
English
@pigeon__s yeah kindle is the current checkpoint in chatgpt pro
it’s SVG results are worse than 5.5
it’s not that much better overall
English