Sabitlenmiş Tweet
All gemini models in lm arena that was upcoming. Google is cooking
1. Night Whisperer
2. Dreamtides
3. Moonhowler
4. Dragontail
5. Stargazer
6. Shadebrook
7. riverhollow
English
AlLeakWire
855 posts
AlLeakWire
@AlLeakWire
Unearthing the hidden side of Al. Tracking leaks, decoding rumors, and revealing what's next in Artificial Intelligence before it's official
Katılım Nisan 2022
1.4K Takip Edilen462 Takipçiler
@marmaduke091 You have seen the outputs of gemini 3.5 pro and gpt 5.6 can you tell me in details how they performs
English
Why is Gemini Omni so bad?
Demis Hassabis introduced this model "as a step to AGI", but I don't see that at all.
Did DeepMind lost their way?
Lentils@Lentils80
Wtf is Google doing bro... Veo 3 was revolutionary when it first dropped last year but they've been on a downward spiral ever since. This better cost 1 cent per 15 seconds and even then I wouldn't use it.
English
Gemini 3.5 Flash vs GPT-5.5 instant vs Sonnet 4.6
Remember guys. #1 in Finance Agent v2. SOTA performance right here. lol 🤣
Prompt : " 300+140=460 Is this correct? Breakdown? "
Vals AI@ValsAI
Google's Gemini 3.5 Flash is the new #1 model on our Finance Agent benchmark (v2), dethroning GPT-5.5 by six points.
English
AlLeakWire retweetledi
This is getting incredibly frustrating, google is kinda faking up lot of things. They said weekly limits are reset and they 3x ed the limits but after just 1 5 hr cycle It shows 60% limits left for all models which resets in 6 days, this is 3x reduction not increase.
🧵
English
@OfficialLoganK @mercor_ai Improve the tool use agentic cases and creative writing
English
Gemini 3.5 Flash ranks #1 on the APEX-Agents-AA benchmark, outperforming much larger models a whole size above it.
English
@ammaar When it will be available on Android I have pre ordered tell me the date
English
Approved and coming to your iPhone soon :)
Preorder for free now: apps.apple.com/us/app/google-…
English
@OfficialLoganK @clanker_ You are 6 months behind from open ai and anthropic and also shipping very slowly compare to competition
English
@clanker_ no plan on giving up, but feedback welcomed as you try out everything we shipped at IO : )
English
I am not happy with Google, the race now is with Anthropic and OpenAi (we are still waiting for Grok)
1. I expected google to have their own open claw
2. I expected gemini to be as good as Claude in coding
3. And i expected gemini spark to operate the entire phone (android)
English
@OfficialLoganK @MaksimXBT @GoogleAIStudio We need this type of powerful model that can beat gpt 5.6
English
@MaksimXBT @GoogleAIStudio we are also experimenting with on device models!
English
ai studio mobile: build your ideas, wherever you are
coming soon to an app store near you
ai.studio/mobile
English
🚨Gemini web UI update
here is the first look for everyone
> now we can select thinking level
> ui looks so clean and good
> main problem of last update is solve now , user will have slider to decide for reasoning < previously it was set at standard in all case >
English
Its definitely on steroids, its insanely fast at both writing code and testing using chrome extension
Chaos@MahawarYas27492
Holy 3.2/3.5 flash has been rolling out in antigravity and its much faster and also in many cases performing better than 3.1 pro. If this is true, Antigravity will be my go to again. @Angaisb_ @test_tm7873 @SaiNemani1 @ai_for_success @nicdunz @koltregaskes
English
⚡ Gemini 3.2 Flash is already out in Gemini!
You can try it with selecting "thinking" and "canvas". I have tried it, and it's pretty good, let me know how it goes 👀
I don't know if they are going to call it 3.2 or 3.5, but they are definitely going to release it on I/O IMO.
English
@Similarweb How was your calculate the market share of gemini chatgpt and claude
English
Instants was released by Instagram four days ago and has already hit #1 on the US App Store.
English
Gemini 3.2 flash
> xbox controller SVG
Harshith@HarshithLucky3
Google silently routing to Gemini 3.2 flash or 3.5 flash in Gemini web 👀👀 use thinking + canvas, you may get it I am seeing this
Dansk
🚨 Gemini 3.2 Pro one-shotted Windows 98 on @arena.
what actually works:
window drag, minimize, maximize
functional browser inside the OS
tools menu with real interactions > games , calculator , paint , msword , notepad
full taskbar
proper boot/login feel
arena link in the reply — go click around yourself 👇
Chetaslua@chetaslua
🚨 Gemini 3.0 Pro - ecpt checkpoint Holy shit Guys , i want everyone to see this retweet as much as you can to get this to mainstream , i dont ask for this normally All apps work , apple animation , minimize , tools , browser , and everything literally this is the best we can see AI till date @OfficialLoganK you have cooked , i wanna see the world when gemini 3.0 drops officially
English
🚨 Gemini 3.5 Flash on @arena → Minecraft in a single prompt.
Best implementation of minecraft ever seen by any ai model till date for one shot
What actually surprised me:
>zero external assets (no .png textures, no models)
>working block place + break with hit feedback
>procedural textures drawn in code, not loaded
Arena link in reply to verify by yourself
Chetaslua@chetaslua
🚨 Gemini 3.5 Flash checkpoint notes: The old “cool Gemini UI” vibe feels mostly gone. It is more GPTified now. Good: fast, smart for a Flash model. Bad: it keeps filling websites with HUDs/panels, even when the prompt asks for restraint. Biggest issue: prompt adherence. I told it multiple times not to use web, and it still did. Because of that, exact knowledge cutoff is hard to verify confidently.
English
@AlLeakWire Responds blazing fast , I don’t have option to connect as mcp rn …maybe we will see it after I/O 👀
English
