Ever since LLMs became a public-facing tech, I haven't spent a dollar on it. I would mess around with the free models, but quickly discover that the LLM tended to just make stuff up on the spot. If I wanted to pay for a product that pulled answers from thin air, I’d buy a Magic 8-Ball; at least it has the decency to tell me to 'ask again later' instead of making up a fact.

However, something began to shift as the tech got better. Now, I have finally surrendered my cash to an LLM, and if I'm being honest, I think I got really lucky with my timing.

It all started with Gemini 3 putting OpenAI on "code red"

It definitely caught my interest

I started using Google Gemini when its third model shook the AI world so hard that it put OpenAI on code red. Surely, anything that can make the original "king of LLMs" sweat bullets has got to be good. So, I gave Google Gemini 3 a try, and I was really happy with it.

If you haven't used Google Gemini's free tier, you get access to two modes: Fast and Thinking. The Thinking mode was great for breaking down complex topics and performing surface research, albeit you only get a handful of prompts. It felt like the number you get depends on the demand on the servers at the time of use, but I usually squeezed 5-10 Thinking prompts before it was done. Meanwhile, Fast felt perfect for general queries, and while I'm seeing people claim that there are limits on the free tier, I never hit them.

I really came to love the Thinking mode on Gemini 3. If I wanted to do a deep dive in a topic, I could feed it into Gemini and then ask future questions from its results. It got to a point where I would use up all my daily Thinking prompts, wait until the limits were lifted the next day, and then use them all up again. I knew I should probably pay for Gemini to get more Thinking prompts, but at $19.99 a month, I really couldn't justify subscribing to the Google AI Pro tier.

The Google AI Plus tier broke the hesitation

And it only just came out

One day, as I was deciding whether or not I'd take the plunge with Google AI Pro, I noticed something had changed. Google used to offer Google AI Pro for a low price for three months, but the trial has now changed to a free one that lasts one month. That got me thinking; if Google AI Pro had changed, what other tweaks has Google made to its plans?

Sure enough, a week after I started seriously using Gemini 3, Google released a new AI plan. It's called Google AI Plus, and it gives you unlimited Fast prompts, 90 Thinking prompts a day, 30 Pro prompts a day, and 5 Deep Research prompts a month. This felt like the perfect middle ground between Free and Google AI Pro for me; it gave me plenty of prompts I could work with in a day, and sliced the price down to $8 a month, with a 2-month trial at $4 a month. That felt a ton more palatable than the $19.99 entry fee Gemini Pro was asking for. So I bit the bullet and subbed.

Now I'm using Google Gemini 3 like a pro

And it has come in great use

Now that I have access to all the options Google Gemini 3 had to offer, I started picking the right weapons for the job. If I want a summary or a simple yes/no answer to a question, I'd use Thinking mode. Pro is much better at churning through really complicated topics, and while I've not given Deep Research a proper try, it does give you a little summary of what it'll look up, which you can tweak. Once you have the battle plans ready, you set Gemini off into the wilds and wait for 10-15 minutes to generate a report. Once the heavy thinking is done, I switch back to Fast to ask basic questions about the data.

I've found that Google Gemini 3 has been an excellent all-rounder tool. I like using it to ask questions, gather trivia, and summarise long documents into an easy-to-digest format. And while my past self is scorning that I'm paying for AI, I honestly think it was a great idea.

Google Gemini 3 has been a great asset in my arsenal

As someone who used to hate the idea of paying for AI, Google Gemini 3 struck a chord with me. It lets me adjust how intense I want it to be when researching, it ticks all the use-cases I have for it, and it now comes in at an excellent entry-level price point. And honestly, as much as I was a stickler before, I don't see myself unsubbing anytime soon.