ChatGPT, Google Gemini & Claude’s BIGGEST Updates From This Week! (AI News Update)

Author: Rob The AI Guy · Watch on YouTube

Video authored by the channel shown, not by IANIA; all rights belong to the creator. Transcript and summary generated automatically; the original-language content is preserved.

Original language: English

Original video: ChatGPT, Google Gemini & Claude’s BIGGEST Updates From This Week! (AI News Update) · Channel: Rob The AI Guy · 10:42 · Watch on YouTube

Topics: AI agents, Task and workflow automation, Prompt and context engineering

Summary

The video reviews updates to ChatGPT, Google Gemini, and Claude, focusing on new models, integrations, and workflow features. ChatGPT updates include a decision model, custom MCP apps, expanded voice capabilities, and Chrome access. Gemini introduces the Gemini 4 Argon model, skills and connectors, and a feature for checking whether media was made with Google AI. Claude adds connections to Google Workspace and tools for working with documents, spreadsheets, and slides.

Key points

ChatGPT’s API adds GPT 61 ultra fast, which the presenter says generates tokens eight times faster than GPT 61 soul, at a higher cost.
A ChatGPT decision model offers predicate, choice, and score functions for evaluating statements, selecting options, and scoring inputs.
ChatGPT plugins can build custom MCP apps, demonstrated with VidIQ features for analyzing a YouTube channel and finding breakout videos.
The GPT live 1.1 voice model adds support for several languages and can interact with files, plugins, and tools in voice mode.
With the ChatGPT Chrome extension installed and updated, ChatGPT can access Chrome tabs and take actions on websites and files.
The presenter promotes My Prompt Buddy for optimizing prompts and saving reusable ones across ChatGPT, Claude, and Gemini.
Google Gemini adds the Gemini 4 Argon model, along with skills, connectors, and a discovery area for additional tools.
Gemini’s Verify AI skill, powered by SynthID, can check whether images or videos were made with Google AI, though it may not identify other AI tools.
Claude can connect with Google Drive, Gmail, and Google Calendar, and can work within Google Docs, Sheets, and Slides.
Claude also offers native tools to create documents, slides, designs, and code without using Google Workspace apps.

Chapters

Start	Chapter	Summary
00:00	ChatGPT API, decision model, and MCP apps	The presenter covers faster API generation, ChatGPT’s decision functions, and custom plugin-powered MCP apps.
03:13	ChatGPT voice and Chrome integration	The video explains the new voice model’s language support and how ChatGPT can interact with Chrome.
04:51	Prompt Buddy promotion	The presenter discusses prompt saving and optimization and recommends the Prompt Buddy tool.
06:27	Gemini 4 Argon and Verify AI	The presenter demonstrates Gemini’s new model, skills, connectors, and feature for checking media made with Google AI.
09:09	Claude and Google Workspace	The video shows Claude connecting to Google services and working with documents, spreadsheets, slides, and native creation tools.

Quotes

this is a way better decision maker than an LLM. — 01:05

this makes it so you don’t have it to type to Chat GPT work anymore — 03:45

you can now tag at Chrome — 04:17

this image was not made with Google AI — 08:36

Claude works now wherever you work. — 10:12

Concepts and entities

ChatGPT, OpenAI, decision model, Jev, MCP, VidIQ, GPT-5.1, GPT-6 Astra, Gemini, Google, SynthID, Claude, Google Drive, Google Docs, Google Sheets, Google Slides, Prompt Buddy, Google Chrome

Full transcript (original language)

00:00 Chat GPT, Google Gemini, and Claude have all released a bunch of brand new upgrades this week that you’ve probably missed because you can now do things like tag Chrome inside of Chat GPT. Google launched a brand new model, Gemini 4, which is way better than any other model that they’ve ever had, and Claude also released a bunch of brand new upgrades. And by the end of this video, you’re going to know what all these are, how you can use them, and how they’re going to help you automate your work. So, the first new Chat GPT update that we need to talk about is now from inside of the API, you’re going to see that you’re going to be able to access GPT 61 ultra fast. Basically, this

00:34 creates tokens eight times faster than just GPT 61 soul. It’s going to cost more money, but it allows you to get way more done way faster. In addition to that, also inside of here, we come over into decisions, we could see that this can now actually go through and there is a decision model from inside of Chat GPT. This is something that Jev released to the world a few weeks ago, but now Chat GPT has one, or rather Open AI has one. And when we could see from right here, we could do things with predicate, we could do things with choice, and we

01:05 could also do things with score. So, under predicate right here, this returns the probability that a statement is going to be true from 0 to 100%. If we come over to choice right here, this selects a choice and returns probabilities for each option. And if we come over to score right here, this evaluates the input against the ordered levels. Now, basically, what this allows to happen is one, this is a way better decision maker than an LLM. So, a Chat GPT model. Two, this is way cheaper and way faster. In fact, it’s almost instant, and the use cases that this unlocks are pretty insane, and I haven’t

01:38 heard anybody talk about this yet. In addition to that, now from inside of Chat GPT, if you have a plugin or your favorite plugins, they’re going to be able to build out their own custom MCP apps. If I come over into these three dots right here, I could see that VidIQ has actually built out one of these already. And basically, what this allows is for a customized experience that’s going to be powered by Chat GPT work. We could see right here, this actually has access to my YouTube channel. It sees all of my videos, it sees all of my breakout videos right here, and actually finds breakout videos from inside of my

02:11 niche. In addition to that, I could come over to keywords right here, and this basically prompts me with different things that I could do with Chat GPT, like find the best keywords for a specific video, or summarize a video, or plan a scheduled upload, or analyze my channel, or find breakout videos. Basically, if you have a Chat GPT plugin, you’re going to want to make sure that you’re creating an MCP app because it just allows for your users to be able to discover things way more easily. And I hope way more of my plugins actually roll this type of stuff out. If we interact with anything from

02:41 inside of here, basically what this then does is this opens up a Chat GPT work window right here where we could actually come over here, and we don’t actually see the prompt that this passed along. But if we scroll down, we could see all of the actual work that this has done, all the data that this has actually taken into consideration, and this has created something that is pretty cool here that otherwise wouldn’t have existed, or I might not have known about. Now, the next update that I did want to talk about quickly is we could see Chat GPT has now released a brand new voice model. The brand new model is going to be GPT live 1.1, and we could see that it supports a bunch of new

03:13 voices. It supports voices in French, in German, in Hindi, in Mexican Spanish, and Korean, and in Japanese. In addition to that, it got rid of that constant checking that this would do when you were trying to do work with it inside of Chat GPT work. And we could see that this is now live inside of dots, inside of codex, and inside of Chat GPT. This is a pretty big upgrade to Chat GPT voice because now, when we’re inside of Chat GPT work and we turn on voice mode, this will actually be able to interact with our files, this will be able to

03:45 interact with our plugins, and this makes it so you don’t have it to type to Chat GPT work anymore in order to get it to be able to access your files, your plugins, or your different tools, it can do all of it from directly inside of here in voice mode, which I absolutely love. In fact, if we come over here and click on voice mode right here, we could see that this now opens this up and we’ll be able to access this just how we would be able to normally. Then that final new ChatGPT update I wanted to make sure that you were aware of is now you can use your Google Chrome from inside of ChatGPT and you could use

04:17 ChatGPT from inside of Google Chrome. So we could see right here, you can now tag at Chrome and say something like I’m finally ready to buy a new sofa, which one of the ones that I’m shopping with and we could see we have all these different tabs open, should we actually buy? And then guess what this is able to do. This is able from just inside of a ChatGPT chat to actually go through our Chrome, go through all of our different tabs, and from inside of ChatGPT work, you could actually get this to go through and interact with your files, download things for you, fill out forms for you, take action on websites for you

04:51 all on your actual computer because it’s able to access it now if you have the ChatGPT Chrome extension installed, you have it up to date on the latest version, this will be able to do all of this. Now before we get into the rest of those new updates, I did want to share with you something that ChatGPT still can’t do, which is that you still from inside of ChatGPT can’t save prompts that you use over and over again and it doesn’t optimize your prompts for you. And one thing that you’re going to notice if you’re using GPT-6 Astra or ChatGPT’s latest models is that the prompt that you give it drastically

05:24 changes the output you get. And that is exactly why I use this My Prompt Buddy Prompt Optimizer right here. This allows you to insert all of your different prompts right here from a standard prompt, reasoning prompt, deep research, AI video, AI image, you could put in your current prompt, your ideal output. And as we could see from my history right here, this thing goes through and actually changes your prompt so that you get a professional prompt right here that is much better. In addition to that, if we come over into shortcuts right here, this is where you can actually save prompts that you reuse over and over again. So, you can simply just copy, paste it directly in here,

05:56 and you don’t have to do anything other than that, and you don’t even have to leave ChatGPT in order to do that. It works with ChatGPT, Claude, and Gemini, so it works with all the LLMs, and this is a no-brainer tool that I use every day. That’s exactly why I’d strongly suggest you go to the pinned comment below and get started with my Prompt Buddy today. I use this tool literally every single day. It’s compatible with Gemini, ChatGPT, and Claude, and it’s going to make your life way easier. The best part about it is if you go to the pinned comment below, you could get started with a free trial today to test it out for yourself. So, change number one I wanted to cover when it comes to

06:27 Google Gemini is now we’re going to see that you can now access Gemini 4 from directly inside of here. So, now when you come over to your model selector, we’re going to see that we have 3.5 flashlight, 3.8 flash, 3.1 pro, and 4 Argon. I would strongly suggest you are using 4 Argon. In addition to that, when we now actually come over into more right here and click on customize, we could see that we can now add skills to Google Gemini. We could add connectors, and they added in a ton more to Google Gemini. In addition to that, if we click

06:58 on custom right here, we can actually add in custom connectors, and if we come over to discover, this will ask you to discover different skills, different business essentials, different plugins that you could actually use from inside of here, which is pretty cool. Now, what makes Gemini 4 Argon special is that this is going to be Google Gemini’s new frontier model right here. So, we could see exactly what this is really good at. If you were analyzing a complex business report, if you were creating a deck, if you were building a 3D model, Argon brings frontier intelligence to your everyday workflows, which allows you to

07:31 turn your ambitious ideas into real projects and outputs that you could use. And so, everybody always asks, where does this actually compare? Let’s look at this. So, we could see that on a a of things, this is better than Claude Opus 5.5, Claude Fable 5.1, and GPD6 Astra. In fact, the only model that is slightly better at this at some things is going to be GPD6 Astra. So, if you were trying to automate things, if you’re trying to do agent decoding, if you’re trying to do machine learning engineering, science, and math, doing things with long context, computer use, multimodal understanding, this blows the other

08:05 models out of the water. On top of that, Google Gemini has released a brand new feature that actually allows you to go through and be able to do at. So, we if we do at right here, we could tag all of our different skills, we could tag all of our different connectors, and we could even use this verify AI skill right here, which is pretty cool. What this actually allows you to do is upload an image, you could upload a video. For example, I’m actually going to upload this screenshot that I took right here. If we click into this, we could see what this screenshot looks like. And what I’m going to get this to do is tell us

08:36 whether or not it’s AI. I’m going to say, “Was AI made to create this?” And then what this will actually go through and do is use verify AI, which is powered by SynthID, and we could see this image was not made with Google AI, but they weren’t able to confirm if other AI tools were used to actually create it. Now, this wasn’t created with AI, so this actually got this correct, and that’s pretty cool. Now, if we contrast that, for example, let’s say that we come over here, and let’s say that we grab this image right here of me, which looks absolutely terrible, but if we take a screenshot of this, we come

09:09 over into new chat, and we actually upload this right here, we could say, “Is this AI?” And then this will actually go through and correctly identify whether or not it’s AI. So, yes, this was generated or edited using Google AI. So, this can now allow you to actually figure out if images or videos were generated with AI. Now, the next update I wanted to show you is going to be to Claude. So, we can now see that we could actually go through and now connect our Google Drive suite, our Gmail, our Google Calendar to Claude.

09:40 Now, if we come over here and click on connect to all, this is basically going to unlock the capability right here where we could come over into extensions and we can actually use Claude right here. So, we can now see that there is Claude for Docs where we could go through and we could ask it actually make changes to this document and boom, it made those changes. Same exact thing works for inside of Google Sheets right here where we could go through and we could actually get this to do our actual work from Google Sheets using Claude. This is absolutely incredible. It also works with Google Slides right here

10:12 where we could go through, we can access our connectors, we could access our plugins, we could access our different skills right here and this is able to go through and connect everything and create anything that we actually want. So, Claude works now wherever you work. On top of that, as I’m sure you’ve already seen, if we actually come over here and we click on plus, we will also be able to go through and actually be able to create different things here. We could get this to create documents, we could get this to create slides, we

10:42 could get this to create whatever we want. In fact, we come over into design right here, we could see slides, design, design and code base. We could get it to create all of these different things and guess what? This is not going to use Google Sheets, Google Slides, Google Docs. This is going to use the Claude native tools. Now, if you enjoyed this video and you want to learn about all of ChatGPT’s latest updates that I didn’t cover in this video, I would strongly suggest you check out this video right here. I’ll see you over there.