I Used Google's Gemini 3.0 for a Week—Here's What Actually Impressed Me

Google launched Gemini 3.0 last Monday, and I've been putting it through its paces ever since. Not with fancy benchmarks or theoretical questions—I used it for the kind of stuff we actually do every day. Writing emails, planning meals, coding, research. The usual. Here's what I found, good and bad.

First Impressions: Speed and Interface

The first thing I noticed is how fast it is. Gemini 3.0 responds almost instantly, even for complex queries. Google redesigned the interface too—cleaner, less cluttered. You can now upload images, PDFs, and even videos for analysis. I threw a 20-page PDF at it, and it summarized it in seconds. That's impressive. But the real test is whether the answers are actually useful.

Writing Assistance: Better Than ChatGPT?

I asked Gemini to draft a professional email for a project proposal. The tone was spot-on—polite but direct. Then I asked ChatGPT 5.0 the same thing. ChatGPT's version was warmer, more human. Gemini's felt a bit robotic. For creative writing, like a short story, Gemini struggled. The prose was stiff. ChatGPT clearly wins for creativity. But for factual writing—like summarizing news articles—Gemini is more accurate. It hallucinates less. I checked a few facts, and Gemini got them right 9 times out of 10. ChatGPT made up a source once.

Coding: A Real Developer's Take

I'm not a professional coder, but I dabble in Python for data analysis. I asked both to write a script to scrape a website and clean the data. Gemini 3.0 gave me a clean, working script on the first try. ChatGPT's version had a bug—it used an outdated library. Gemini also explained the code better, with comments that actually made sense. For coding, I'd pick Gemini. It's more reliable for current syntax.

Research and Reasoning

I asked a tricky question: 'What were the economic impacts of the 2024 Tokyo earthquake?' Gemini gave a detailed answer with specific GDP figures and recovery timelines. It cited sources from the IMF and Japanese government. ChatGPT's answer was more general. For research tasks, Gemini's access to real-time data (via Google Search) gives it an edge. But it sometimes includes irrelevant details. I had to sift through a few paragraphs to find what I needed.

Multimodal Features: The Killer App

This is where Gemini 3.0 shines. I uploaded a photo of my messy fridge and asked it to suggest a meal plan using what I had. It identified all the items—even the half-eaten jar of pickles—and gave me three recipes. That's wild. I also uploaded a graph from a scientific paper, and it explained the trends in plain English. ChatGPT can do this too, but Gemini is faster and more accurate with images.

Where It Falls Short

Gemini 3.0 is not perfect. It struggles with sarcasm and humor. I asked it to tell me a joke, and it was painfully unfunny. It also has a tendency to be overly cautious—it refused to answer a question about DIY home repairs because it 'might be unsafe.' That's annoying. And the free tier is limited to 50 queries per day. I hit that limit by lunchtime. The pro version costs $30/month, which is steep compared to ChatGPT Plus at $20.

Final Thoughts

So is Gemini 3.0 worth it? If you're a researcher, coder, or someone who relies on accurate, fast information, yes. The multimodal features alone make it useful. But if you need a creative assistant or a chatbot that can banter, ChatGPT is still better. I'm using both for now. Gemini for work, ChatGPT for fun. That seems like the smartest move.