92
Carbon produces impressive 3D and visual work, often beating rivals
Gemini 4 Carbon: Early user impressions
Do early public impressions judge Gemini 4 Carbon positively or negatively?
- Point change
- −9 pts
- Current state
- Super smart
- Status
- Running
More monitor statisticsLess monitor statistics
- Point change
- −9 pts
- Current state
- Super smart
- Status
- Running
Early impressions
preevia.com25%Not great
0%Mixed
75%Super smart
0%
75%
0%
90%
67%
0%
0%
90%
25%
75%
Updated Nov 8Since Oct 28
Where people stand
A sample of the public voices behind the data.
@dylan522pOfficial apology to Google from me, they cooked with Carbon.
It's spikier in terms of domains and capabilities, but it's really good in a few areas.
@galnagliI'm really enjoying Gemini 4 Carbon, I always find myself asking it how it figures out its way to complex chains - never had so many novel discoveries while bug hunting
If you’re still not
@chetaslua🚨 Gemini Carbon 4 genuinely is so good in threejs
this is a google model we are praising , no laziness at all
now it think like really thinks before output sometimes 8-10 min
This is ser
@pankajkumar_devGemini 4 Carbon (Hydrogen) : My Impressions/Review
- Gemini 4 Carbon H is a new checkpoint available in Arena.
- I tested this model in Arena, and it’s quite good at Three.js. I would say t
@Mr_nexp🚨 Gemini 4 Carbon Announced
Google is taking on GPT-6 Astra, Opus 5.5 and Fable 5.1 and the results are insane.
Google is back at the frontier, and I genuinely didn't expect Carbon to go
@LuminaBenchGemini 4 carbon looks crazy
This is what the people wanted https://t.co/FpwpKIR1Ft
@Mr_SalioGemini 4 Carbon output
This model is seriously good at 3D and Three.js. It handles some pretty challenging tasks with ease this took around 3h
This is the new checkpoint appeared on Arena
@patmeenanCarbon is the first Gemini model I've used that doesn't lie about doing the work when I give it a hard coding problem. Even Argon was fabricating results.
I still have trust issues, but it 
@AnikaCodesSeen dozens of AI game demos this year, but this 30-minute Minecraft run on Gemini 4 Carbon actually made me stop and watch.
Right now Fable 5.5 is the only other model playing in that leag
@Haleeeemahhgemini 4 carbon 3D test
it did really good here and it was fast!
testing games tomorrow lets see what this model can do :) https://t.co/qoe60QuwCj
@JamesMalsawmGemini 4 Carbon is clearly the strongest model in the world.
If you were OpenAI, Anthropic or SpaceX, would you even continue competing with Google? https://t.co/j8Oksnc6OJ
@TimJayasgoogle just killed opus 5.5
gemini-4-carbon official benchmark just dropped and it did what openai couldn't do. it's just better!!
@amoufarekGemini 4 Carbon is a very good model and I look forward to more people getting access to it.
Combined with Google's super power (cost-efficient scale) and you get the best perf/cost ratio f
@dummerspast39Gemini 4 CARBON created Ye's head out of voxels, and it's by FAR the only good one out of ALL models.
💧
No other model can create a proportional head OR anything resembling Kanye West, but
@astropol0Gemini 4 Carbon is literally Google's most strategic drop yet...
surpassed Argon by 9 full points in a two week sprint to grab the #1 spot on Artificial Analysis and the long horizon agent
@SPAC89🔴 Gemini 4 Carbon is now the best model on the Artificial Analysis Intelligence Index v5.0, even beating Opus with a score of 59 vs 58! This is such a huge achievement for Google, I've been 
@alexgetmancom🚨 Gemini 4 Carbon is out
According to synthetic benchmarks, it’s now the best AI model in the world. The Artificial Analysis Intelligence Index seems to confirm that.
But will it actually 
@OctoberryGemini Carbon (through Arena) did this
It took about 30 minutes, but the result is kinda impressive to me lol, also I think it's got the beak shape well too
What do you think? https://t.c
@worksaltCarbon is benchmaxxed bullshit, just like Argon. I have already watched youtube videos of what Gemini 4 Carbon does in 3D applications, and visually it is easy to see that it is not as good 
@AI_ScreeningGemini 4 Carbon Checkpoint in Arena vs. Claude Opus 5.5
>Carbon is really strong competition for Opus 5.5
>Really excited to see Fable 5.5 beat both https://t.co/ees2StQnyyNot greatMixedSuper smart
The views shaping the conversation
Recurring ideas, summarized across multiple conversations.
Not greatMixedSuper smart
73
Carbon ranks at or near top on synthetic benchmarks
92
Carbon is seen as a reliable, honest coding and reasoning model
20
Some users call Carbon overhyped, benchmark-gaming, and weaker in 3D
85
Carbon wins cost comparisons against Opus 5.5 on creative tasks
59
Carbon's results are spiky, strong in some areas and weak in others
These are collected experiences, not a survey of all users. Errors and post miscalculations are possible.
20% signal support · 693 conversations