From Gmail to video: what Google’s Gemini does today, what each plan costs and when ChatGPT or Claude may be the better choice.
CTRL + ALT News | Artificial Intelligence | Updated: 30 October 2026
Transparency note: this text was written with the help of Gemini itself, at the site’s request, and then checked by us against the official pages of Google, OpenAI and Anthropic. We removed from the original whatever had no support on those pages, such as the forum and video sources. The site also uses Claude, made by Anthropic. Because the question in the headline is a matter of opinion, we show the data and where Gemini is not the best.
Artificial intelligence is no longer a laboratory curiosity. Today it summarises emails, searches the internet, helps you write, creates images and writes code. Among the big assistants, Google’s Gemini holds a card the other two don’t: it already lives inside services many people use every day.
Is that enough to make it “the best AI”? The honest answer is that it depends on what you do. Here is what Gemini offers, what it costs and where it falls short.
1. What is Gemini and how does it work?
Gemini is Google’s family of AI models and also the name of the app and assistant that uses them. As with ChatGPT, it helps to separate the two: the model is the engine, and the app is the complete vehicle, with its interface, tools and connections. If these terms are new to you, see our explainer on LLMs, AI agents and agentic AI.
The most advanced model Google lists today is Gemini 4 Argon, which the company describes as its flagship reasoning model. According to the developer documentation, it accepts text, images, video, audio and PDFs and has a context window of about 1 million tokens (1,048,576). Put simply, that is how much text the model can consider at once.
Accepting several kinds of file is called multimodality. In practice you can send a photo, an audio clip or a PDF and ask for an analysis, without converting anything first.
Source: Google — Gemini API: models.
2. The biggest argument: integration with Google’s services
If there is one reason to consider Gemini, this is it. It works inside Google services you may already use. According to the subscriptions page, the AI Plus plan includes Gemini in Gmail and Vids, and the AI Pro and Ultra plans include Gemini in Gmail, Docs and Vids.
The page also says Deep Research can use the context of your Gmail, Drive and Chat, and that Gemini Live can search Gmail and Drive, check Google Calendar and add items to Google Tasks or Keep.
An illustrative example: you ask, “summarise this week’s emails from client X and find a free slot in my calendar for us to talk”. It is the kind of task where living inside Google helps, because nothing needs copying into another tool. We did not test that request, and what is available varies by plan and account settings.
For businesses, Google introduced on 8 October 2026 the “Gemini agent”, an agent for work that can schedule tasks, react to events and be reached through channels such as Google Workspace, Microsoft 365 and Slack. Google’s text mentions Gemini inside Gmail, Drive, Docs, Slides, Sheets, Chat and Calendar, but gives no plans, prices or availability dates.
The advantage is not only the model: it is where the model sits. If you don’t use Google’s services, you lose much of this argument.
Sources: Google — Gemini subscriptions and Google Cloud — Welcome to Gemini at Work 2026.
3. Search and Deep Research: when you need sources
AI models don’t always know recent facts, which is why web search matters. Gemini has Deep Research, a feature for longer investigations. According to Google, it browses hundreds of websites and delivers multi-page reports in minutes, in more than 45 languages. The subscriptions page lists Deep Research on the free plan already.
For developers, Google describes a Deep Research agent that plans and carries out multi-step research “across hundreds of sources”, but it is still in preview.
One necessary caution: the original text we received said Gemini is “unbeatable” at finding recent information “without hallucinating out-of-date data”. No official page supports that. Being connected to a search engine does not prevent errors. Always check the sources a report cites, especially on health, money or law.
Source: Google — Gemini subscriptions.
4. Images, video and voice in one place
Another strength is bringing media together in one app. According to the subscriptions page, the free plan includes image generation and editing, Gemini Live (voice conversation) and Canvas, a space for creating and editing. Video generation appears on the paid plans, and the release notes of 19 May 2026 describe Gemini Omni for creating video, for subscribers, with availability that varies by region.
Gemini Live lets you talk by voice and, according to the page, can remember details across sessions. Limits on images and videos per plan are not given as numbers.
By comparison, OpenAI’s pricing page also lists image generation (limited on the free plan and unlimited on Pro). On Claude’s page we did not find image generation, which does not prove it doesn’t exist. For anyone who wants images, video and voice in one place, Gemini and ChatGPT are well placed.
5. October news: Gemini 4 Argon, Skills and Antigravity
Gemini’s release notes show a fast pace of change. These are the latest items we could confirm:
- 13 October 2026: Gemini 4 Argon began rolling out globally to Pro and Ultra subscribers. The same day, Google Antigravity (an agent-based coding tool) became built into the Gemini app for macOS, and the “Customise” feature began connecting more than 40 apps, most of them in US English.
- 30 September 2026: Skills arrived in chat for repetitive tasks. Google says they will replace Gems, and that Gems will be migrated.
- 10 September 2026: the Gemini app for Windows arrived, and it can connect to Gmail and Drive.
The practical consequence: features change name and place quickly, and what holds today may change within weeks.
Source: Google — Gemini release notes.
6. Gemini at work: from chat to agent
Just as ChatGPT has Work and Codex, Google has been taking Gemini from chat to carrying out tasks. The “Gemini agent” presented on 8 October is described as a universal agent for work that runs in the cloud, uses tools and can create sub-agents. Google also talks about identity controls, permissions and audit logs.
For programming, Antigravity comes with the Pro and Ultra plans, with entry-level and higher limits, according to the subscriptions page. In an independent test of coding agents by Artificial Analysis, Gemini 4 Carbon with Antigravity scored 59 points, behind Claude Code with Opus 5.5, which scored 62.
A detail we declare for transparency: Google’s text says the Gemini agent orchestrates Gemini models and also “Claude models from Anthropic”.
7. How much does Gemini cost? Is it worth paying?
| Plan | Price on the page | What changes |
|---|---|---|
| Free | £0 | Flash-Lite model, images, Deep Research, Live, Canvas, Gems and Notebook; 15 GB of storage |
| AI Plus | £4.49 a month | About 2x the free plan’s usage, Flash model, video, Gemini in Gmail and Vids, 400 GB |
| AI Pro | £18.99 a month | Gemini 4 Argon, 4x the free plan’s usage, Gemini in Gmail, Docs and Vids, Antigravity with entry-level limits, 5 TB |
| AI Ultra | From £79.99 a month | Highest access (up to 20x the Pro plan’s usage on the top tiers), 20 TB or more |
Usage limits are calculated from the complexity of your request, the features you use and the length of the conversation, and they refresh every five hours until you reach a weekly limit. The page doesn’t say how many messages that means. Check the price for your region when you subscribe.
Which plan makes sense? To try it out, the free plan already offers a lot. If you mainly want Gemini in Gmail, look at AI Plus, the cheapest paid entry point of the three assistants (ChatGPT Go costs £7 and Claude Pro $20). Pro makes sense if you want the most advanced model and Gemini in Docs.
Source: Google — Gemini subscriptions.
8. Gemini, ChatGPT or Claude: which should you choose?
| Gemini | ChatGPT | Claude | |
|---|---|---|---|
| Paid entry price | AI Plus: £4.49 | Go: £7 | Pro: $20 |
| Mid-range plan | AI Pro: £18.99 | Plus: £20 | Max: from $100 |
| Context window (per the pages) | 1 million tokens on Pro | 256,000 (Plus) and 400,000 (Pro) in reasoning mode | Up to 1 million on all plans |
| Email and documents integration | Gmail, Docs, Vids | Connected apps (Work) | We did not check |
| Images and video | Yes (images on the free plan; video on paid plans) | Images; unlimited on Pro | We did not find any |
For the equivalent ChatGPT news, see our piece on ChatGPT’s intelligent interface.
Prices are in different currencies because each page uses its own; compare with care. The ChatGPT and Claude rows come from the pricing pages read on 30 October 2026.
### What do independent tests say?
On 28 October 2026, Artificial Analysis put Gemini 4 Carbon at the top of its intelligence index v5 with 59 points, followed by Claude Opus 5.5 (58) and GPT-6 Astra (57). Two caveats: Gemini 4 Carbon had restricted access on the date of the analysis, and the name “Carbon” does not appear on the Google pages we read. And the index measures models in specific configurations, not the experience of the apps.
None of these figures says a single assistant is the best for everyone.
9. When Gemini is the best choice (and when it isn’t)
Gemini tends to be a great choice if:
- You live inside Google. Gmail, Drive, Docs and Calendar are its strongest point.
- You want images, video and voice in one app, with much of it available on the free plan.
- You want to start without paying or pay little: the free plan is broad and AI Plus is the cheapest paid plan of the three.
It is worth looking at the alternatives if:
- Your focus is programming with agents. In Artificial Analysis’s coding-agent test, Claude Code with Opus 5.5 came out ahead.
- You don’t use Google’s services. Without them, Gemini’s biggest advantage shrinks.
- You value ChatGPT’s set of features, such as Work and Codex, which we describe in our article on whether ChatGPT is the best AI in 2026.
The original text said Claude delivers a “more human text texture” and that ChatGPT has “hyper-realistic” voice and a “giant ecosystem” of mini-apps. We found none of that on the official pages or in an independent test, so we removed those lines. To see how each one does on everyday tasks, see our comparison Claude, ChatGPT or Gemini: which AI should you choose.
10. What Gemini doesn’t do perfectly yet
Like any generative AI, Gemini can make mistakes, misread a request or present a source that doesn’t support what it says. That applies to search answers, image and file analysis, and any long task.
### Privacy needs attention
Gemini’s overview page (which dates from 2024) says you can stop your future conversations being used to improve Google’s machine-learning technologies by turning off the “Gemini Apps activity” setting. Business plans have their own rules: Google says that on them the data is not used to train models. We did not open Gemini’s privacy page, so check these settings before sending sensitive information.
Use AI to extend your ability to work and learn, not to replace your own judgement. For medical, legal and financial decisions, confirm with reliable sources and qualified professionals.
11. Frequently asked questions about Gemini
### Is Gemini the best artificial intelligence of 2026?
It is one of the most complete options, especially for people who use Google’s services. But there is no proof that it is better at every task. In programming with agents, for example, Claude leads an independent index.
### Is Gemini free?
There is a free plan, with the Flash-Lite model, image generation, Deep Research, Gemini Live and other features, within usage limits. The paid plans widen those limits and add features.
### What is the difference between Gemini and Gemini 4 Argon?
Gemini is the family of models and the app. Gemini 4 Argon is the most advanced model Google lists today, available on the Pro and Ultra plans.
### Does Gemini read my Gmail?
On the plans that include Gemini in Gmail, it works inside the service, according to your settings. Check the account’s permissions and activity settings before using it.
### Can Gemini create videos?
Yes, on the paid plans, according to the subscriptions page, with limits and availability that vary by region.
Conclusion: the best AI is the one that fits your day
Gemini has a strong argument: it is already where many people work, and it brings images, video, voice and search together in one app with a broad free plan. That makes it one of the most practical choices for anyone who lives in Google.
That doesn’t mean it is the best for everyone. For programming with agents, Claude comes out ahead in an independent test, and ChatGPT brings together features such as Work and Codex. The ideal choice depends on what you do, your budget and the results you get in practice.
Which AI turns your ideas into useful results best? For many people who use Gmail and Docs every day, Gemini deserves to be among the first options to try.
What we could not confirm
Plans, prices and features come from official Google pages with no publication date, which change often. The Gemini overview page dates from 2024 and the release notes were read only for the most recent part. We did not open the privacy page. We did not test Gemini. Gemini 4 Carbon does not appear on the Google pages we read, and Artificial Analysis’s results depend on each model’s configuration.
### Sources and references
Leave a Reply