ChatGPT vs Claude vs Gemini: Which $20 AI Plan Is Worth It?

0:00 Chat GPT, Claude, and Gemini all cost $20 a month, and yet what you actually get for that $20 is really different depending on which one you pick. I paid for all three for a full month and ran them on the same tasks to see where each one pulls ahead and where it falls apart. By the end of this, you'll know

0:15 exactly which of these three plans deserves your $20 based on the way you actually use AI. So, I put all these three AI models through the same set of tests built around the things people use these tools for. That means making video and images. Which of them has the smartest model behind it? How good each

0:30 one is at writing and everyday questions? How well it plugs into your other apps? Whether it can do real work on the files sitting on your own computer? How it handles code? And how much you get to use before it cuts you off? Every one of those categories has a winner. So, as we go, I'll call who

0:43 takes each round, and by the end you'll watch the wins stack up into a clear pattern. I'll take them one at a time starting where the three are furthest apart, which is making media. If you want to make video, Google's plan is the only one of the three that even does it. Inside the Gemini app, you describe a

0:58 scene in plain words and it builds the video for you using Veo, which is Google's video model. And you can take it further by turning a still photo into a video or dropping yourself into a scene with the built-in templates. The other two aren't even in the competition. Chat GPT used to make video

1:12 through Sora, but OpenAI shut the consumer version down in early 2026. So, there's no longer a video button anywhere in Chat GPT Plus. And Claude has never done video at all because it's built for writing, analysis, and work rather than making media. The one catch is that the full highest quality Veo

1:28 sits on Google's pricier Ultra plan. So, on the $20 tier, you get the lighter, faster version with tighter limits. But it's still real video generation the other two don't offer. So, the video round goes to Google, and it isn't close because it's the only plan of the three that can hand [music] you a finished

1:42 video. Images are closer, and this is again a round where Claude sits out. Claude can look at an image you upload and describe it in detail, but it can't create one. So, the moment you need a picture made, Claude is not an option. That leaves Chat GPT and Gemini, and both are strong. With Chat GPT, you

1:57 describe what you want and it generates it. And its real strength is getting text inside an image right and letting you edit the result with a comment. So, you change one detail and it adjusts that part instead of starting over. Gemini generates and edits images, too, through Google's Nano Banana model. And

2:11 the quality is just as good with the same trick of uploading a photo and having it changed however you describe. For text-heavy work like a poster or an infographic, ChatGPT has the slight edge right now. Though for realistic edits to an actual photo, Gemini's Nano Banana is right there with it and often a bit

2:26 faster. So, for everyday image work, you'd be happy with either. So, the image round is ChatGPT's by a small margin with Gemini just behind and Claude not in it at all. That's two media rounds gone and Claude has sat out both, which already tells you something about where its strengths really sit.

2:41 Because the moment we stop talking about making media and start talking about raw intelligence, everything changes. Every one of these plans is built around a different AI model. So, the fair test is which of those models is the smartest [music] right now. The newest models from each are GPT 5.5 inside ChatGPT,

2:57 Gemini 3.1 Pro inside Google's plan, and Claude [music] Opus 4.8 inside Claude Pro. And in practice, they're remarkably close, separated by only a few points on the main independent intelligence ranking. So, whichever one you land on, you're getting a genuinely capable model. Right at the top of that overall

3:12 ranking today sits Claude's Opus 4.8 just ahead of GPT 5.5 with Gemini 3.1 Pro a step behind the two of them. But the overall number hides the more useful part, which is that each one is strongest [music] in its own area. Gemini 3.1 Pro is the cheapest of the three to run by a clear margin, which

3:29 makes it the value pick of the group. GPT 5.5 is the strongest at fully autonomous tasks, where you set it going and leave it to work on its own without checking in, which is exactly what OpenAI built it for. And Claude Opus 4.8 leads on coding, and it tends to stay closest to the facts on long answers,

3:46 inventing less when a question gets complicated. So, on raw intelligence, the round goes to Claude, but only just. And the truer takeaway is that all three are very close, with the smartest one for you being whichever is strongest at the thing you care [music] about most. And keep that in mind because it sets up

4:00 the next category, where raw intelligence matters far less than tone. And by the way, since every round has a different winner, I put the whole scorecard into a free cheat sheet who takes each category and which plan to buy based on the way you actually use AI. You can grab it in the description

4:14 below. Day-to-day though, what you actually notice is how each one writes and how it talks to you. Start with plain everyday chat, the asking questions and getting answers you'll spend the most time on. And here the three feel different in personality. Gemini tends to give you the most

4:27 structured, organized answers. ChatGPT sits in the middle with a clear shape you can also adjust a little in the settings. Claude leans the other way, often writing longer, more flowing answers that read like a person explaining something rather than a list, which some people love and others find

4:41 harder to scan. There's no clean winner on everyday chat because it's the most personal thing in the whole comparison. So I'd call this round a draw and tell you to trust your own taste. Though if I had to send a total beginner somewhere for easy, readable answers, I'd [music] point them at Gemini. Proper writing

4:55 though, a real document rather than a quick reply, does have a winner and it splits in a way worth understanding. On the quality of the words themselves, [music] Claude is the strongest of the three because it's writing is the most natural and human, the least likely to read like a robot produced it. It can

5:09 also learn your voice since you upload something you've already written through its skills feature and it matches your tone. So what it [music] writes back sounds more like you than like generic AI. Claude also lets you keep all of this organized in projects, which are saved workspaces where you keep the

5:23 files, notes and instructions for one ongoing piece of work. So it remembers the context every time you come back instead of you having to set it all up again. But the words are only half of writing because the other half is the workspace where you shape them. And that's where Gemini takes over. With

5:36 Gemini, you can open a tool called canvas and have it write something for you. It opens in a side panel that works almost like a word processor, so you can edit the text directly, change headings [music] and make things bold. And you can also just ask it to rework the writing and it updates the draft right

5:50 there. Chat GPT also lets you edit drafts directly inside cleaner writing blocks, so the workflow feels much more hands-on than Claude's simpler document [music] panel. So, this round splits two ways with the best writing going to Claude and the best [music] place to actually work on a document going to

6:04 Gemini. So, pick on the quality of the words if that's [music] what matters to you and pick Gemini's canvas if you spend your time polishing the document itself. Writing for yourself is one job, but these tools get far more useful the second you let them reach into the other apps your life already runs on. The

6:18 moment you want one of these to plug into your email, your calendar, or your notes, it gets more interesting because all three now reach much further than they used to. Claude has a very wide reach with the big names like Gmail, Drive, and Notion all there. Plus a link to Zapier that connects it out to

6:32 thousands of other apps. Chat GPT is close behind covering [music] a lot of the same big tools, and while Zapier isn't a built-in app there, you can still reach it through Chat GPT's actions. Gemini has caught up more than people realize because alongside the full Google Suite, it now connects to

6:46 outside apps like Canva and GitHub. So, the connectors round goes to Claude on sheer breadth, though the margin is thinner than it was since all three can now connect to plenty of tools outside their own. The one exception is if you already run everything through Google because then Gemini's tight link to

6:59 Gmail, Docs, and Drive is the smoothest of the three for that one setup. However, connecting to your apps is one level and the step up from there is an agent that can sit on your computer and do the work itself. A real agent is one that works through a task on its own, and on these three plans only Claude

7:14 gives you one that runs on your actual computer for everyday work. Claude's version is called Cowork and it lives in the Claude desktop app where you point it at a folder on your machine and it reads, edits, and creates files right there running locally on your computer rather than off in the cloud. People use

7:28 it for the dull useful stuff, sorting invoices, organizing a project folder, drafting documents straight from their own material, and it's one of the strongest tools available for real knowledge work, meaning basically anything that isn't pure coding. It even has a feature called dispatch that turns

7:42 your phone into a remote control for it, so you can send a task from the mobile app like check my calendar and unread emails or send me a summary before my first meeting, and Cowork runs it on your desktop while you're away, all on the same $20 pro plan. ChatGPT does have an agent called Codex, but it's built

7:58 for something different because it's made for coding and runs in the cloud on your code repositories rather than working through the everyday files on your computer. So, Codex is useful, just not a general assistant for your own files the way Cowork is. Gemini's answer is Google anti-gravity, which is free

8:14 and does run on the files on your computer, but just like Codex, it's a coding tool for developers. So, it works on code and not the everyday files that Cowork handles. So, the round for real work on your own files is Claude's, and it stands alone because the nearest things from ChatGPT and Gemini are both

8:29 coding tools, while Cowork is the only one made for the everyday files on your computer. That covers general work, but narrow it down to writing code specifically and the order changes again. Coding is the one area I'll keep tight since I'm not a developer myself, but the picture among people who are is

8:42 fairly clear. If you're writing code, Claude and ChatGPT are the two clear leaders, and each has a dedicated coding agent built for it. Claude has Claude Code, a tool developers point at their projects to plan changes, edit files, and run commands. And on the benchmarks, Claude's model scores highest on real

8:57 coding tasks. ChatGPT answers with Codex, the same cloud agent from a moment ago, and it's almost as strong, so you won't go wrong with either. Most developers still give Claude a slight edge for how consistent and reviewable its code tends to be, with ChatGPT a close second. Gemini can code, too, and

9:12 the one place it really stands out is front-end design since it's the best of the three at turning a layout into a good-looking website or app. So, if you're building something where the look matters to you, Gemini's the one worth a try, even if for the heavier back-end work, it's still third. So, the coding

9:26 round goes to Claude by a slim margin, with ChatGPT close behind. And this is the category where the two of them leave Gemini furthest back. That's the work side handled, and there's one last thing that decides the whole value question, and it's got nothing to do with the chatbot itself. Every plan has a

9:40 standout extra, but one of them beats the other two so completely, it nearly ends the comparison on its own. Start with Claude, whose standout is a tool called Claude design, where you build interactive prototypes, slide decks, and design mockups from a couple of prompts. And it's one of my favorite things it

9:54 does. ChatGPT's real strength outside of chat is its image work and its templates, plus that Codex agent for developers. So, most of its value stays inside the core product. But, this category belongs to Google, and it isn't close. For the same $20, Google AI Pro doesn't just hand you Gemini, it adds a

10:10 whole pile of things on top. You get higher limits on Notebook LM, one of the best research tools Google has ever made, plus AI music generation, the Flow Studio for video, and Gemini built straight into Gmail, Docs, and Sheets to help you write and organize. Then come the perks with nothing to do with the

10:25 chatbot at all, because you also get 5 TB of cloud storage across [music] Gmail, Drive, and Photos, the Google Home Premium plan you'd otherwise pay for on its own, and $10 a month in Google Cloud credits. You can even share the plan with up to five other people, [music] though it's the main AI perks

10:39 they get rather than every single extra. So, the extras round is Google's, and it isn't close because it's the only plan that includes a stack of other tools and perks for the same price. All that value counts for little, though, the moment you run out of usage halfway through the

10:51 day, which is the last thing left to check. So, the final round comes down to usage, which is simply how much each one lets you do before it cuts you off and makes you wait for a reset. ChatGPT is the easiest to [music] understand because OpenAI gives you clear numbers, somewhere around 160 messages every 3

11:07 hours on its strongest [music] model, with that window resetting on a 3-hour cycle. Claude and Gemini work differently because instead of counting messages, they weigh how heavy your requests are. So, the more you pile into a single prompt, the faster you use it up. And both reset on a 5-hour window

11:20 with a weekly cap on top of that. In my experience, Claude runs out the fastest of the three by a clear margin. And [music] when you hit its limit, it simply stops you until the reset, with no fallback model. It's also worth knowing that co-work and dispatch use the same allowance as your normal chats.

11:34 So heavy work there eats into the usage you'd want for everything else. Chat GPT and Gemini are gentler because hitting their limit doesn't lock you out. It just drops you to a weaker model so you can keep going. And in my experience, Gemini lasts the longest of the three because its limits are compute-based. So

11:49 as long as you're not piling huge requests into one prompt, everyday use barely touches them. That leaves the usage round split three ways with the most predictable limits on Chat GPT, the most forgiving on Gemini, and Claude the one most likely to run out on you, especially if you lean on it hard. So

12:03 when you add it up, Google AI Pro is the value and range pick, the easy call if you want the widest spread for your $20 since it's the only one that does video and piles on a stack of extras while being the hardest to run out. [music] Chat GPT Plus is the steady, no surprises choice, the one to get if

12:18 images matter and you want limits you can predict with the sharpest image tools and the clearest everyday allowance. And Claude Pro is the work pick, worth it despite the tighter limits if your time goes on writing, real tasks on your own files, or coding because it writes the most naturally and

12:32 its co-work agent is brilliant for getting actual work done. The best $20 plan really comes down to the work you're doing. And whichever of these three you pick, you'll only get your $20 worth if you actually know how to use it. That's the whole point of my school community, getting you from total

12:45 beginner to really good with these tools. So when you're ready, go ahead and join with the link in the description and I'll see you inside. Thank you for watching and I'll see you on the next one.