Inmersion retweeted
Introducing Gemini 4 Hydrogen, an extremely capable and fast model built to bring Gemini 4 to everyone!
Available to all tiers of Google AI Studio, Gemini API, Gemini for business, the Gemini app, Google Search, and paid Google AI Pro + Ultra subscribers in Google Antigravity.
Inmersion retweeted
Gemini 4 Krypton: November 25th for Cyber Defenders, December 8th for public availability, according to Reddit.
Inmersion retweeted
Google is preparing an XHIGH reasoning tier for Gemini!!!
And this time, it's showing up in Google's own GenAI SDK.
'ThinkingLevel.XHIGH'
Meanwhile, Gemini 4 Carbon currently offers:
• Low
• Medium
• High
Inmersion retweeted
Gemini 4 Hydrogen is already being tested internally at Google, according to Business Insider.
Its also on Arena, and many people have tested it, including me.
According to a Google engineer, its a very capable, fast and efficient model.
So the launch may not be far away, as Google seems satisfied with the model.
Inmersion retweeted
🚨 Fable 5.2 flew Superman through the whole history of computer graphics ,
1972 white blocks → 2026 path tracing < pure javascript , no images , no models , no video >
Inmersion retweeted
Gemini 4 carbon is now available for Pro users on AI mode/Search
unfortunately not in any other environments at the moment
Inmersion retweeted
Gemini 4 Carbon is live today. I've been testing it for a few weeks.
Here's a 3D model Carbon made for me of the new landscraper office, using only 150 PDFs from the council
Gemini 4 Carbon is rolling out today. It’s built for sustained, long-horizon tasks, reasoning deeply through complex software engineering, cybersecurity, and scientific discovery.
Coming globally via the Gemini API in @GoogleAIStudio and Gemini for business, and for Ultra subscribers in @antigravity, the @GeminiApp, and AI Mode in Search.
Inmersion retweeted
The reality right now 🤣
We've given all Pro, Max, Ultra, and Team users one free full reset, across all models and limits. Use it anytime before December 10 at 2:00PM PT.
We recently fixed a few small issues that used more limits than intended, and the Fable resets didn't do much for those of you who'd already hit your weekly limits for all models. While these issues didn't affect everyone, we want to make up for the trouble.
You'll find it in Settings > Usage, or at clau.de/reset.
Inmersion retweeted
carbon is the most efficient google model
asked it to make me smash kart clone and model did really well with this one:
> took about 1 hour 30 mins
> used 2M tokens worth ~$2.15 only
team really cooked harder with this one as it's cheaper and efficient too
we get frontier model performance at fraction of a cost but it's super slow even slower then argon
here's what it made, did it cook or nah?
fable 5.5 vs gemini 4 carbon
tested both models with same prompt at highest reasoning available and results came out kinda same
> fable took 60 mins and cost ~$18
> carbon took same time as fable and cost ~$6.2
carbon is super slow just like fable and even way slower then argon, no idea why but it did really well on high while fable was tested on xhigh
which model did better here?
Inmersion retweeted
I’m blown away. Gemini 4 Carbon, what are you cooking? 🤯
It built this GTA 6-style volume slider in 12 minutes for $1.22
