I hesitate even to say this (it causes me a significant amount of embarrassment and regret), but I used to spend $200 monthly on ChatGPT Pro.

I guess maybe that's the cost of doing business (I have found quite the niche being an AI writer, mainly focused on OpenAI products, so it's my duty), but it was beginning to be too much.

You may be surprised (or not) to find out that I canceled it yesterday.

Why, you may ask? Well, it's all because of a little model called Gemini 2.5 Pro.

Google is Winning the AI Race

It's funny, at any given time in the past year, any of the following players could have taken the spot of "most intelligent LLM":

• Anthropic (Claude 3.7)
• OpenAI (OpenAI o1 and o3)
• Deepseek (R1)
• Grok (Grok 3)
• Google (Gemini 2.0 Pro)

However, there is now an undisputed king, and it's none other than Google.

Gemini 2.5 Pro — The Lodown

Google's bleeding-edge reasoning model is far and above the competition…

…okay, it's not that much better than o3-mini… but there's a huge catch.

It's FREE.

That's right, to have access to o3-mini (the runner-up OpenAI model), you have to pay the $200/month fee to OpenAI. With Google Gemini, you get 50 messages free each day.

It makes you wonder, is there even an argument to stay with OpenAI at this point? Well there's more than just intelligence when it comes to LLMs, so I figured I'd do a breakdown of the two.

Gemini vs. GPT — The Breakdown

In this section, I want to do a blow-by-blow examination of the two AI titans so you can make the call for yourself if you want to switch. I even made a little scorecard based on the following categories:

• Intelligence + Context Window (5 points)
• Cost + Usage Limits (4 points)
• User Experience (3 points)
• Canvas (2 points)
• Voice Mode (2 points)
• Image Generation (2 points)
• Extra Features (1 point)

Intelligence + Context Window

I think it's quite apparent here that Gemini takes the cake, given the intelligence charts. In my time using both, I simply find Gemini 2.5 Pro to be better at following instructions, and even more so, it's great at inferring things. When using LLMs, I typically find myself having to baby them along, giving them every single modicum of detail. With Gemini, I save so much time.

When it comes to context window size (how much text we can insert into each model before it starts "forgetting"), Gemini wins by a factor of 8, coming in at over 1,000,000 tokens (~750,000 words) vs. ChatGPT's 128,000… and you only get the full 128 if you pay $200/month.

This category goes to Gemini.

Scores:

Gemini: 5 Points

ChatGPT: 0 Points

Cost + Usage Limits

Welp, clearly there is no argument here. While both providers offer a free tier and a paid tier, Google's offering is much stronger.

Google gives 50 messages of Gemini 2.5 Pro per day, and infinite messages of Gemini 2.0 Flash, and Gemini 2.0 Flash Thinking, both competent models.

OpenAI, on the other hand, doesn't even give infinite access to GPT-4o (let alone o1 and o3-mini) for plus users ($20/month).

This category clearly goes to Google.

Scores:

Gemini: 9 Points

ChatGPT: 0 Points

User Experience

Here's the thing. I love ChatGPT's UX. In fact, I love it so much that it's probably why I haven't switched to another LLM provider sooner.

However, Gemini's is also quite nice:

While it does hold a candle to OpenAI, Google's LLM interface does lack a few critical features:

1. No search bar for chats
2. A less intelligent memory and custom instructions section
3. No projects

Because of these reasons, I give this category to OpenAI.

Scores:

Gemini: 9 Points

ChatGPT: 3 Points

Canvas

Both providers give us a handy canvas mode for easy interactive editing and coding with the LLM:

Both seem to have nearly identical functionality, although each remains a bit buggy from time to time, especially when the context gets large. This one is a tie!

Scores:

Gemini: 11 Points

ChatGPT: 5 Points

Voice Mode

Both models come with capable voice models, allowing for interactive and engaging speech interactions.

However, I find that OpenAI beats out Google on two metrics here:

1. ChatGPT's voice model is simply higher quality, more realistic, and a better more human experience.
2. In order to use voice mode with Google, you have to go to Google AI Studio, a completely separate webpage from where you'd normally use Gemini.

For these reasons, I give this category to OpenAI.

Scores:

Gemini: 11 Points

ChatGPT: 7 Points

Image Generation

OpenAI's image generation models are all the rage right now. It's incredibly good at following directions and can create some really creative and interesting compositions, not to mention its impressive capabilities with text.

However, Google isn't too far behind with their Imagen 3 Model.

For technical image creation (such as thumbnails, style transfer from existing images, and image editing), OpenAI clearly wins.

The following test was a one-shot example, with one identical request for each. Here is the prompt I used:

Make an image that shows an artistic rendering of the painting inspired by the creation of adam, except it's a cyborg's arm reaching for a human. Masterpiece quality.

Google Imagen 3:

OpenAI 4o Image Generation:

A brutal defeat of Google by OpenAI.

Scores:

Gemini: 11 Points

ChatGPT: 9 Points

Extra Features

For this last section, I just want to review some nice features from both providers. Depending on your workflow, any of these could be a deal-maker or breaker.

1. GPT Store — (OpenAI has this, Google doesn't)
2. Deep Research — (Both providers have this)
3. Desktop microphone input — (Google has this, OpenAI somehow still doesn't have it… this makes it much easier to do stream-of-consciousness interaction with an LLM)
4. Projects — I mentioned this earlier, but it's such a valuable feature that I'll mention it again. Google, you gotta catch up here!
5. Tasks — ChatGPT can schedule tasks for you and send you notifications. Gemini can't! … yet.
6. Temporary Chats — Don't want a mega-corporation training their models on your data? Well, OpenAI provides a self-deleting, non-trainable chat experience for this. Google doesn't!

I may be missing some features here, but I think OpenAI takes this one.

Scores:

Gemini: 11 Points

ChatGPT: 10 Points

The Verdict

Well, it was closer than I thought, I'll admit. At the end of the day, Google squeaks out the win here. Granted, my naive scoring system isn't the most balanced, but that's how I rate the value of LLM adjacent features.

If OpenAI is true to their word, o4-mini and o3 are on the horizon, which may blow Gemini 2.5 Pro out of the water. If that's the case, I might think about switching back.

Until then, I'm a Gemini dude.

Thanks for reading!

-Jordan