Google Launches Gemini 3, Its 'Most Intelligent' AI Model Yet

Google released Gemini 3 on Tuesday, launching its latest AI model with a breakthrough score of 1501 Elo on the LMArena Leaderboard alongside state-of-the-art performance across multiple benchmarks including 91.9% on GPQA Diamond for PhD-level reasoning and 37.5% on Humanity's Last Exam without tool usage. The model is available starting today in the Gemini app, AI Mode in Search for Google AI Pro, Google AI Studio, Vertex AI and the newly launched Google Antigravity agentic development platform. Third-party platforms including Cursor, GitHub, JetBrains, Manus, and Replit are also gaining access.

Separately, Google said AI Overviews now have 2 billion users every month. Gemini app has topped 650 million users per month.

Yeah...

[Rosco P. Coltrane]

Higher quality slop and approaching pink slip.

Will it help you avoid paying money to the google?

[shanen]

Not an unreasonable FP, but now you can reply to yourself to clarify what you mean. Perhaps under a more thoughtful and substantive Subject? I'm too happily retired to care about your apparent focus, and I don't see how to make any more jokes under the current conditions so thoroughly dominated by Poe's Law, so I'm switching to a practical topic:

The google is nagging me, even leaning on me, to buy some storage. But "Who steals my data, steals trash." I'm quite sure that my stored data is strongly dominated

Google, your user counts are fraudulent

[ebunga]

You don't get to count users of your AI overviews feature when you didn't give them a choice.

Re:

[Anonymous Coward]

Gemini user counts do not count the search.

Re:

[thegarbz]

The number isn't Gemini users, it's AI Overviews. They are *ONLY* counting the search.

Re: Google, your user counts are fraudulent

[devslash0]

They're being a bit of a count about it.

Does this mean it'll stop sucking?

[Slartibartfast]

I've really, really been unimpressed with Gemini. I'll use almost anything else.

Also -- completely tangential -- can I just say how nice it is that Slashdot still exists? Of course, it's WAY different than Chips 'n Dips and Redhat IPO days, but still, it's fun to come here and feel 1998 wash over me...

Re:

[ebunga]

It will never stop sucking. Our only hope is the AI winter.

Re:

[garcia]

I found GP2.5 to be great at academic-style research and writing; it was absolutely awful at writing code. So; I would tell it to plan some thing for me and write it in a way that could be used by another agent (Claude Code) to build the code to do the thing. In this way, it has been great! I haven't yet attempted it with 3.

That said, I found GP3.0's page to be hilarious:

It demonstrates PhD-level reasoning with top scores on Humanityâ(TM)s Last Exam (37.5% without the usage of any tools) and GPQA Diamo

Re:

[dsgrntlxmply]

If there's anything that touts PhD-level reasoning and writing, it's a recipe book.

Example recipe book with a PhD author: Modernist Cuisine [a.co]

Re:

[jythie]

I am not even sure what 'PhD level reasoning' would even mean. PhDs are specializations, meaning detailed about some narrow topic. Reasoning has nothing to do with it unless you are in a specific field that involves theory on, well, reasoning.

Re:

[Hadlock]

Gemini is one of if not the top model, but I intentionally avoid using it as the other models are within a few percent as good as it, and I'm not at all eager to feed the google machine this time around. Now that google search is irrelevant I have no incentive to "help" google along. Goodbye and good riddance.

Re: Does this mean it'll stop sucking?

[wgoodman]

Ponies was peak /.

Re: AI is biased against anything not Big Pharma

[paul_engr]

Maybe chiropractors are mostly pseudoscientific quacks and frauds? Not all of them, but enough where they're clearly not self policing

Re:

[ForkInMe]

Unfortunately, I think the ones you think are "good" would be the ones eliminated if they were self policing- they as a whole, believe (or profess to believe) some wacky shit...

Re:

[spaceman375]

I personally know five chiropractors. Four of them are serious about medical science in their practices. The fifth I can't vouch for, she's too creepy to visit professionally. Do you actually know and have experience with even one chiropractor?

Re:

[jythie]

Meaning well and being serious about their field is not mutually exclusive with being pseudoscience quacks. They are usually not con artists, they really believe in what they are doing.. but just like their customers it is a belief rooted in conspirtal thinking and selective bias.

Re:

[spaceman375]

a belief rooted in conspirtal thinking and selective bias.

I repeat: Do you actually know and have experience with even one chiropractor? Vague handwaving about a stereotype does not make a valid point. You are just regurgitating the hot air others have farted in your direction. Despite the dogpile or derision it receives, chiropractic practice is based on the scientific method. I've never met a licensed chiropractor who follows aromatherapy, homeopathy, "Life Coach" or other "I can't get a degree so I'll

Re:

[ForkInMe]

I know two actually, one is the son of a good friend of mine and believes he can cure/prevent asthma, cancer, and pretty much anything else. The second is the brother in-law (ex now I guess) of another friend of mine - He was/is a real believer; when he got testicular cancer, he put off real treatment for almost a year while trying his pseudoscience and almost died - then relented and got chemo and radiation and lived.

Re:

[jythie]

Yeah, but just like other snake oil, people are REALLY committed to the mythology and lash out at anything that contradicts it. When you can paint any disagreement as part of a vast evil conspiracy, critical thinking is just a punchline.

Socialization

[nycsubway]

How does this help me (or anyone) socialize? It doesn't help bring people together, it drives them apart into their own little silos with the AI being their only gateway outside of that silo.

We live in a time of great contrast

[MpVpRb]

Research continues on developing AI that will help us solve previously intractable problems in science, engineering, medicine, etc
The general public misuses it to create slop and scams
Investors irrationally pour billions into old ideas
Hypemongers, pundits and futurists invent fantastic fiction

Re:

[jythie]

Eh, one of the big open secrets in the field is AI researchers (as opposed to tech bros) tend to be pretty open about how AI should never be used for anything where getting an answer right is important. The hussle crowd loves selling them as solutions to science, engineering, and medicine, but meaningless slop is all it is really good for and the only place where people are willing to actually pay (since those tend to be use cases that make money while having no legal liability)

Re:

[spacepimp]

the only way for the revenue models to meet the expected returns of investors means that AI needs to be able to help everyone everyday with most everything. Limiting it to intractable problems in science and medicine means the AI companies will never get the returns they need to justify the trillions in build out.

Duh

[devslash0]

Yes, everyone's products are always latest and greatest. Always "the most X yet". You know why? Because they are the latest.

Ask Gemini what version it is.

[Going_Digital]

Apparently google doesn't want you to know if you are using the 2.5 or version 3, what are they tying to hide? It simply refuses to answer the question, how completely dumb.