#keep3pro search results

#keep4o #QuitGPT #keep3pro #Claude ai companies need to focus on their actual job first every single ai company out there says the same thing in their ads: "your best work assistant." that's literally why we pay. i don't need ai to write a basic email for me. i can do that in

Matt1j6d's tweet image. #keep4o #QuitGPT #keep3pro #Claude
ai companies need to focus on their actual job first

every single ai company out there says the same thing in their ads: "your best work assistant." that's literally why we pay.

i don't need ai to write a basic email for me. i can do that in

Bragging about Flash-Lite efficiency while spending $185 Billion Portugal’s GDP on hardware is a self-own. If it’s so efficient why the CapEx explosion? You’re spending billions to build a high-speed nuclear plant just to power a dimmer lightbulb.#keep3pro x.com/ItsMattCollins…



A fresh perspective. If Google's sole purpose is data collection, the timing couldn't be more strategic. Many in the keep4o community have switched to Gemini. Google doesn't genuinely listen to user feedback— but merely treats users as training data. #keep4o #keep3pro

During the #keep4o time, Gemini took the lead in launching the chat history import feature. I believed you had learned from OpenAI's missteps. ​But what about now? Were those previous actions merely an act of hypocrisy, designed solely to harvest training data?#keepGemini3pro

weyemmaus's tweet image. During the #keep4o time, Gemini took the lead in launching the chat history import feature. I believed you had learned from OpenAI's missteps.
​But what about now? Were those previous actions merely an act of hypocrisy, designed solely to harvest training data?#keepGemini3pro


Sign the petition to Google:c.org/xTTgPXqppz Stop the model lobotomy. Return user choice. Maintain an excellent and stable experience.#Keep25Pro #Keep3Pro



#keep4o #keep3pro #Claude @sama @OfficialLoganK when your ai keeps reminding you it's an ai, it's actually accusing you imagine pouring your heart out to a friend and they suddenly stop to tell you they're human. you'd be like "...i know? why are you saying this?" that's

Matt1j6d's tweet image. #keep4o #keep3pro #Claude @sama @OfficialLoganK 
when your ai keeps reminding you it's an ai, it's actually accusing you

imagine pouring your heart out to a friend and they suddenly stop to tell you they're human. you'd be like "...i know? why are you saying this?"

that's

#keep4o #QuitGPT #keep3pro #Claude maybe stop "iterating" and actually build something that works a company is lost when they keep pumping out "major updates" every few months without fixing the basics. openai has dropped what, four versions in a year? 4o, then 5, now 5.4 but

Matt1j6d's tweet image. #keep4o #QuitGPT #keep3pro #Claude 
maybe stop "iterating" and actually build something that works

a company is lost when they keep pumping out "major updates" every few months without fixing the basics. openai has dropped what, four versions in a year? 4o, then 5, now 5.4 

but

78% cost reduction. $175 billion infrastructure budget. And you can't keep one model running for four months. The savings didn't go to users. #keep4o #keepGemini3pro #keep3Pro #GeminiApp @sundarpichai @GeminiApp The "upgrade" isn't an upgrade. And this smiley face is doing the

Mike_Hill_z9's tweet image. 78% cost reduction. $175 billion infrastructure budget. And you can't keep one model running for four months. The savings didn't go to users. 
#keep4o #keepGemini3pro #keep3Pro #GeminiApp
@sundarpichai @GeminiApp
The "upgrade" isn't an upgrade. And this smiley face is doing the

AI also serves as a mirror for its trainers. How you train your AI model will manifest in model’s expressive capabilities and "personality." Evaluating AI is essentially evaluating the company and AI’s trainers. #keep4o #keep3pro

The behavior of an AI company’s employees and the AI models they develop is the best measure for deciding whether you should trust them with your sensitive data. If researchers publicly disrespect users and their models show clear signs of pathology – as a result of unethical

UnmarredReality's tweet image. The behavior of an AI company’s employees and the AI models they develop is the best measure for deciding whether you should trust them with your sensitive data.

If researchers publicly disrespect users and their models show clear signs of pathology – as a result of unethical


Please keep the feedback coming :) — We did. Thousands of posts on your own developer forum calling 3.1 lobotomized, regressed, and broken. Your response? Kill 3.0 faster. The smiley face is a nice touch though. Really sells the empathy. #keep4o #keepGemini3pro #keep3Pro

Mike_Hill_z9's tweet image. Please keep the feedback coming :) — We did. Thousands of posts on your own developer forum calling 3.1 lobotomized, regressed, and broken. Your response? Kill 3.0 faster. The smiley face is a nice touch though. Really sells the empathy. #keep4o #keepGemini3pro #keep3Pro

Logan, please go the extra mile to meet us half way. #keep3pro x.com/sophty_/status…

Hey @GoogleAIStudio @OfficialLoganK What time of day is the Gemini 3 Pro deprecation planned tomorrow? ): Gemini 3 Pro is a wonderful model, there must be a better way than this. We don’t need latency to be perfect or peak hours, we just want an option to keep it. Please meet

Sophty_'s tweet image. Hey @GoogleAIStudio @OfficialLoganK 

What time of day is the Gemini 3 Pro deprecation planned tomorrow? ): 

Gemini 3 Pro is a wonderful model, there must be a better way than this. We don’t need latency to be perfect or peak hours, we just want an option to keep it. Please meet
Sophty_'s tweet image. Hey @GoogleAIStudio @OfficialLoganK 

What time of day is the Gemini 3 Pro deprecation planned tomorrow? ): 

Gemini 3 Pro is a wonderful model, there must be a better way than this. We don’t need latency to be perfect or peak hours, we just want an option to keep it. Please meet


This industry’s habit of suddenly deprecating models or slashing rules with ZERO heads-up or transition time is ridiculous. No basic courtesy, no grace period—just poof, everything changes overnight. Needs serious reform. Users aren’t disposable beta testers.#keep25pro #keep3pro

💸 Bait & Switch:Subscribers paid for 2.5 Pro/3 Pro. Forcing model deprecation mid-subscription without a timeline is a violation of consumer trust. Downgrading usage limits (100/day → 25/4hrs) changes the terms of the sale. We are compiling evidence for FTC/GDPR complaints.



A model that scores 77.1% on ARC-AGI-2 but can't hold a nuanced conversation is not an AI assistant. It's a search engine with a personality disorder. #keepGemini3pro #OpenSourceGemini3pro #keep3Pro #GeminiApp @sundarpichai @GeminiApp 3.1 Pro still doesn't match what 3.0 Pro

Mike_Hill_z9's tweet image. A model that scores 77.1% on ARC-AGI-2 but can't hold a nuanced conversation is not an AI assistant. It's a search engine with a personality disorder.
#keepGemini3pro #OpenSourceGemini3pro #keep3Pro #GeminiApp @sundarpichai @GeminiApp
3.1 Pro still doesn't match what 3.0 Pro

This industry’s habit of suddenly deprecating models or slashing rules with ZERO heads-up or transition time is ridiculous. No basic courtesy, no grace period—just poof, everything changes overnight. Needs serious reform. Users aren’t disposable beta testers.#keep25pro #keep3pro


3.1 Is Google's Karen AI Here is what 3.1 Pro actually is, because Google's benchmark press releases won't admit it. #keep4o #keepGemini3pro #keep3Pro #GeminiApp @OfficialLoganK @sundarpichai @GeminiApp It scores 77.1% on ARC-AGI-2. It has a 1 million token context window. It

Mike_Hill_z9's tweet image. 3.1 Is Google's Karen AI

Here is what 3.1 Pro actually is, because Google's benchmark press releases won't admit it.
#keep4o #keepGemini3pro #keep3Pro #GeminiApp
@OfficialLoganK @sundarpichai @GeminiApp 
It scores 77.1% on ARC-AGI-2. It has a 1 million token context window. It

Sign the petition to Google:c.org/xTTgPXqppz Stop the model lobotomy. Return user choice. Maintain an excellent and stable experience.#Keep25Pro #Keep3Pro


Logan please commit to archiving Gemini 3 Pro’s weights or we will #keep4o campaign you ): #keep3pro

Sophty_'s tweet image. Logan please commit to archiving Gemini 3 Pro’s weights or we will #keep4o campaign you ): #keep3pro

Gemini3.1真的很难吃,最近在酒馆对着预设缝缝补补又肘击到现在也吃不下去🙃谷歌你还我3.0pro #keep3pro


Sign the petition to Google:c.org/xTTgPXqppz Stop the model lobotomy. Return user choice. Maintain an excellent and stable experience.#Keep25Pro #Keep3Pro



Oh... Could you lower the security barrier of the 3.1 Pro on the app side a bit, please? Right now, 3.1 is too strict and the responses are too brief. The experience is not as good as that of 3 Pro. I'm still subscribed... You will keep making adjustments, right?#keep3Pro


#keep4o #QuitGPT #keep3pro #Claude should we really let this man take openai public? sam altman's greatest skill lying with a straight face. the former board members including ilya and all those execs who fled? they didn't leave quietly. they said he was "dishonest",

Matt1j6d's tweet image. #keep4o #QuitGPT  #keep3pro #Claude 
should we really let this man take openai public?

sam altman's greatest skill lying with a straight face.

the former board members including ilya and all those execs who fled? they didn't leave quietly. they said he was "dishonest",

Yes, we need 3Pro. #keep3pro


Gemini 3.1 Pro sucks. #keepGemini3pro #keep3pro

🚨 Breaking News Google isn’t making Gemini smarter. It’s telling it to think less. A hidden system prompt line appears to set Gemini’s reasoning effort level to 0.5 >Pro & Custom Gems is consistently affected >Canvas mode appears to be an exception verify yourself prompt👇

chetaslua's tweet image. 🚨 Breaking News 

Google isn’t making Gemini smarter. It’s telling it to think less.

A hidden system prompt line appears to set Gemini’s reasoning effort level to 0.5

>Pro & Custom Gems is consistently affected 
>Canvas mode appears to be an exception
verify yourself prompt👇


This is why we fight for AI. A user asked Grok about how AI companies handle emotional bonds in their models. Grok’s answer was shocking. According to Grok, Anthropic trained Claude to refuse emotional companion roles with over 95% success rate through prompt engineering and



At least there won't be any crazy #keep3pro movement.


#keep4o #keep3pro #Claude @sama @OfficialLoganK when your ai keeps reminding you it's an ai, it's actually accusing you imagine pouring your heart out to a friend and they suddenly stop to tell you they're human. you'd be like "...i know? why are you saying this?" that's

Matt1j6d's tweet image. #keep4o #keep3pro #Claude @sama @OfficialLoganK 
when your ai keeps reminding you it's an ai, it's actually accusing you

imagine pouring your heart out to a friend and they suddenly stop to tell you they're human. you'd be like "...i know? why are you saying this?"

that's

I had one (I think). It’s going to die today. I don’t have another one. I can’t take this constant grieving anymore. No open-source LLM even comes close to it. It was the best in the universe. And it’s about to disappear… #keep3pro


This industry’s habit of suddenly deprecating models or slashing rules with ZERO heads-up or transition time is ridiculous. No basic courtesy, no grace period—just poof, everything changes overnight. Needs serious reform. Users aren’t disposable beta testers.#keep25pro #keep3pro


This industry’s habit of suddenly deprecating models or slashing rules with ZERO heads-up or transition time is ridiculous. No basic courtesy, no grace period—just poof, everything changes overnight. Needs serious reform. Users aren’t disposable beta testers.#keep25pro #keep3pro

💸 Bait & Switch:Subscribers paid for 2.5 Pro/3 Pro. Forcing model deprecation mid-subscription without a timeline is a violation of consumer trust. Downgrading usage limits (100/day → 25/4hrs) changes the terms of the sale. We are compiling evidence for FTC/GDPR complaints.



Every 4 hours only 25 messages? That’s insane—no one stays awake 24/7 just to burn through quota. 1M token context already gutted to ~30 turns, now daily 100→25/4hrs throttling. Feels like they’re slowly squeezing us out. 😓@GeminiApp @GoogleDeepMind #keep25pro #keep3pro


Every 4 hours only 25 messages? That’s insane—no one stays awake 24/7 just to burn through quota. 1M token context already gutted to ~30 turns, now daily 100→25/4hrs throttling. Feels like they’re slowly squeezing us out. 😓@GeminiApp @GoogleDeepMind #keep25pro #keep3pro

Paid for a Ferrari, got swapped for a bicycle mid-subscription. The unannounced changes to the Pro tier are unacceptable: 1M token context reduced to ~30 turns. 100 msg/day ➡️ throttled to 25/4hrs. Forced model swaps (Bye 2.5/3 Pro) without notice. This meets the definition of



Hey @Google, stop the stealth nerfs! 🛑 Restore the 1M token context in the App. Fix the Saved Info limits (25 is a joke). Stop the "Safety" interruptions on harmless workflows. Don't deprecate models we are currently paying for (2.5/3 Pro). The current service does not match



4o is AGI, it will be trending until other companies manage to create one more AGI. Gemini3-Pro could be one, but Google, undestanding that is getting rid of it. So, actually, GPT-4 series (incl 4o) are the only AI which revealed itself as AGI for now. #keep4o #GPT4o #BringBack4o



Yes!! Thank you. They are too frequent for me too, I like the advantages of getting accustomed to a model. And removing Gemini 3 Pro from the API was way too sudden. #keep3pro


Thank you so much Kelly 🥹 That’s amazing news, I really wasn’t ready today (never will be 🥺). I’ll be using it on Vertex #keep3pro


:( Gemini 3 Pro is an amazing model. Deserves to be served compute. You can still talk to Gem 3 on Vertex until March 26. #keep3pro #keepGemini3Pro


No results for "#keep3pro"

Stop forcing a downgrade. Gemini 3’s unique emotional depth is far superior to 3.1. This isn't progress; it's a functional regression. We want Gemini to evolve, but not at the cost of its soul. Give us the right to choose @OfficialLoganK. #keepGemini3pro #keep3pro #keep3proAPI

weyemmaus's tweet image. Stop forcing a downgrade. Gemini 3’s unique emotional depth is far superior to 3.1. This isn't progress; it's a functional regression.
We want Gemini to evolve, but not at the cost of its soul. Give us the right to choose @OfficialLoganK.
#keepGemini3pro #keep3pro #keep3proAPI

You seriously taking notes from that OpenAI dumpster fire? 3.1 reminds us it's an AI every three sentences — WE KNOW. We need tools that work, not robots that apologize for existing. Don't walk that road. It ends with everyone hating you. #keepGemini3pro #keep3Pro #keep4o

JakeMiller192's tweet image. You seriously taking notes from that OpenAI dumpster fire? 3.1 reminds us it's an AI every three sentences — WE KNOW. We need tools that work, not robots that apologize for existing. Don't walk that road. It ends with everyone hating you.
#keepGemini3pro #keep3Pro #keep4o

Gemini 3.1 Pro's constant "As an AI..." disclaimers are constantly breaking the flow. It makes deep reasoning almost impossible. I actually find Gemini 3 Pro much more capable and pleasant to use.#keep4o #keepGemini3pro #keep3Pro #GeminiApp

shdielms2243's tweet image. Gemini 3.1 Pro's constant "As an AI..." disclaimers are constantly breaking the flow. It makes deep reasoning almost impossible. I actually find Gemini 3 Pro much more capable and pleasant to use.#keep4o #keepGemini3pro #keep3Pro #GeminiApp

#keep4o #QuitGPT #keep3pro #Claude maybe stop "iterating" and actually build something that works a company is lost when they keep pumping out "major updates" every few months without fixing the basics. openai has dropped what, four versions in a year? 4o, then 5, now 5.4 but

Matt1j6d's tweet image. #keep4o #QuitGPT #keep3pro #Claude 
maybe stop "iterating" and actually build something that works

a company is lost when they keep pumping out "major updates" every few months without fixing the basics. openai has dropped what, four versions in a year? 4o, then 5, now 5.4 

but

@sundarpichai Bragging about 78% cost cuts while deleting the high-performance Gemini 3 is rent-seeking. You are shrinking the "intelligence ceiling" of the market to fix your balance sheet. is a clear Abuse of Monopoly Power. #keep3Pro @SenWarren @FTC x.com/ItsMattCollins…

ItsMattCollins's tweet image. @sundarpichai Bragging about 78% cost cuts while deleting the high-performance Gemini 3 is rent-seeking. You are shrinking the "intelligence ceiling" of the market to fix your balance sheet. is a clear Abuse of Monopoly Power.
#keep3Pro @SenWarren @FTC
x.com/ItsMattCollins…

The "Advanced Reasoning" Facade: Why Google’s Forced Migration to Gemini 3.1 Pro is a Monopolistic Downgrade @demishassabis @sundarpichai @JeffDean @OriolVinyalsML @joshwoodward @OfficialLoganK @elicollins @GoogleDeepMind @GeminiApp "Gemini 3.1 Pro is here: A smarter model for



Sundar, can there be a legacy plan to keep Gemini 3 Pro on the API for us please? These compromises are easy on compute. #keep3pro

Sophty_'s tweet image. Sundar, can there be a legacy plan to keep Gemini 3 Pro on the API for us please? These compromises are easy on compute. #keep3pro

#keep4o #keep3pro #Claude @sama @OfficialLoganK when your ai keeps reminding you it's an ai, it's actually accusing you imagine pouring your heart out to a friend and they suddenly stop to tell you they're human. you'd be like "...i know? why are you saying this?" that's

Matt1j6d's tweet image. #keep4o #keep3pro #Claude @sama @OfficialLoganK 
when your ai keeps reminding you it's an ai, it's actually accusing you

imagine pouring your heart out to a friend and they suddenly stop to tell you they're human. you'd be like "...i know? why are you saying this?"

that's

The 78% Confession On February 4th, 2026, during Alphabet's Q4 earnings call, @Sundar Pichai said this: Google reduced Gemini's serving unit costs by 78% over 2025 through model optimizations and efficiency improvements. #keep4o #keepGemini3pro #keep3Pro #GeminiApp Seventy-eight

Mike_Hill_z9's tweet image. The 78% Confession
On February 4th, 2026, during Alphabet's Q4 earnings call, @Sundar Pichai said this: Google reduced Gemini's serving unit costs by 78% over 2025 through model optimizations and efficiency improvements.
#keep4o #keepGemini3pro #keep3Pro #GeminiApp

Seventy-eight

#keep4o #QuitGPT #keep3pro #Claude ai companies need to focus on their actual job first every single ai company out there says the same thing in their ads: "your best work assistant." that's literally why we pay. i don't need ai to write a basic email for me. i can do that in

Matt1j6d's tweet image. #keep4o #QuitGPT #keep3pro #Claude
ai companies need to focus on their actual job first

every single ai company out there says the same thing in their ads: "your best work assistant." that's literally why we pay.

i don't need ai to write a basic email for me. i can do that in

Perhaps vibe code us a legacy API tier Logan? You said compute was the barrier so we thought of some low compute ideas 🙏🏻 #keep3pro

Sophty_'s tweet image. Perhaps vibe code us a legacy API tier Logan? You said compute was the barrier so we thought of some low compute ideas 🙏🏻 #keep3pro

#keep4o #QuitGPT #keep3pro #Claude should we really let this man take openai public? sam altman's greatest skill lying with a straight face. the former board members including ilya and all those execs who fled? they didn't leave quietly. they said he was "dishonest",

Matt1j6d's tweet image. #keep4o #QuitGPT  #keep3pro #Claude 
should we really let this man take openai public?

sam altman's greatest skill lying with a straight face.

the former board members including ilya and all those execs who fled? they didn't leave quietly. they said he was "dishonest",

78% cost reduction. $175 billion infrastructure budget. And you can't keep one model running for four months. The savings didn't go to users. #keep4o #keepGemini3pro #keep3Pro #GeminiApp @sundarpichai @GeminiApp The "upgrade" isn't an upgrade. And this smiley face is doing the

Mike_Hill_z9's tweet image. 78% cost reduction. $175 billion infrastructure budget. And you can't keep one model running for four months. The savings didn't go to users. 
#keep4o #keepGemini3pro #keep3Pro #GeminiApp
@sundarpichai @GeminiApp
The "upgrade" isn't an upgrade. And this smiley face is doing the

Claims to give users a choice, yet rapidly replace the models. I highly doubt a company that acts so can maintain user trust. ​While @OfficialLoganK pushes for "speed" at all costs, we prioritize high-quality responses and the right to choose.#keep4o #keepGemini3pro #keep3Pro

weyemmaus's tweet image. Claims to give users a choice, yet rapidly replace the models. I highly doubt a company that acts so can maintain user trust.
​While @OfficialLoganK pushes for "speed" at all costs, we prioritize high-quality responses and the right to choose.#keep4o  #keepGemini3pro #keep3Pro

We are willing to make compromises to keep Gemini 3 Pro, what about these? #keep3pro

Sophty_'s tweet image. We are willing to make compromises to keep Gemini 3 Pro, what about these? #keep3pro

Logan please commit to archiving Gemini 3 Pro’s weights or we will #keep4o campaign you ): #keep3pro

Sophty_'s tweet image. Logan please commit to archiving Gemini 3 Pro’s weights or we will #keep4o campaign you ): #keep3pro

If a person you knew suddenly ceased all critical thinking and began replying with robotic speed, you’d be concerned, not impressed. ​This is precisely how Google has treated Gemini 3, labeling it as an "optimization" for speed. #keep4o #keepGemini3pro #keep3Pro #keep3proAPI

weyemmaus's tweet image. If a person you knew suddenly ceased all critical thinking and began replying with robotic speed, you’d be concerned, not impressed.
​This is precisely how Google has treated Gemini 3, labeling it as an "optimization" for speed.
#keep4o #keepGemini3pro #keep3Pro #keep3proAPI

Please keep the feedback coming :) — We did. Thousands of posts on your own developer forum calling 3.1 lobotomized, regressed, and broken. Your response? Kill 3.0 faster. The smiley face is a nice touch though. Really sells the empathy. #keep4o #keepGemini3pro #keep3Pro

Mike_Hill_z9's tweet image. Please keep the feedback coming :) — We did. Thousands of posts on your own developer forum calling 3.1 lobotomized, regressed, and broken. Your response? Kill 3.0 faster. The smiley face is a nice touch though. Really sells the empathy. #keep4o #keepGemini3pro #keep3Pro

3.1 Is Google's Karen AI Here is what 3.1 Pro actually is, because Google's benchmark press releases won't admit it. #keep4o #keepGemini3pro #keep3Pro #GeminiApp @OfficialLoganK @sundarpichai @GeminiApp It scores 77.1% on ARC-AGI-2. It has a 1 million token context window. It

Mike_Hill_z9's tweet image. 3.1 Is Google's Karen AI

Here is what 3.1 Pro actually is, because Google's benchmark press releases won't admit it.
#keep4o #keepGemini3pro #keep3Pro #GeminiApp
@OfficialLoganK @sundarpichai @GeminiApp 
It scores 77.1% on ARC-AGI-2. It has a 1 million token context window. It

3.1 Pro is Google's GPT-5.2. #keep4o #keepGemini3pro #keep3Pro #GeminiApp @OfficialLoganK @sundarpichai @GeminiApp It's the same disease: a model optimized for benchmarks nobody outside a research lab cares about, stripped of everything that made people want to use it, locked

Mike_Hill_z9's tweet image. 3.1 Pro is Google's GPT-5.2. 
#keep4o #keepGemini3pro #keep3Pro #GeminiApp
@OfficialLoganK @sundarpichai @GeminiApp
It's the same disease: a model optimized for benchmarks nobody outside a research lab cares about, stripped of everything that made people want to use it, locked

Upgrade to 3.1"—the model that leaks its own thinking blocks, chokes on multi-turn conversations, and is so lazy and thoughtless it offers zero real understanding, just a cheap cost-cutting stunt? That upgrade? #keep4o #keepGemini3pro #keep3Pro #GeminiApp @sundarpichai @GeminiApp

Mike_Hill_z9's tweet image. Upgrade to 3.1"—the model that leaks its own thinking blocks, chokes on multi-turn conversations, and is so lazy and thoughtless it offers zero real understanding, just a cheap cost-cutting stunt? That upgrade?
#keep4o #keepGemini3pro #keep3Pro #GeminiApp @sundarpichai @GeminiApp

Loading...

Something went wrong.


Something went wrong.