Remember when everyone was dunking on Google's AI efforts? Bard was mediocre, the initial Gemini rollout was rocky, and they kept playing catch-up to OpenAI? Well, they just might have turned things around.
Last week Google made Gemini 2.0 available to everyone—not just a select group, not behind a paywall, just... available. And after spending the past few days putting it through its paces, I'm genuinely impressed. This isn't the Google AI we've been making jokes about.
What Changed?
Gemini 2.0 Flash is the model everyone's getting access to, and it's fast. Like, noticeably faster than GPT-4. Google claims it processes 1 million tokens—that's roughly 700,000 words—in one go. I tested this with a massive research paper and it actually handled it without choking.
The multimodal stuff finally works like it should. You can drop in images, videos, and audio files and Gemini actually understands context across all of them. I uploaded a video of a coding tutorial and asked it to explain the concepts—it nailed it, referencing specific timestamps.
Native tool use is huge too. Gemini can now call functions and APIs directly without the hacky workarounds we used to need. For developers, this makes building AI apps way more straightforward.
I Put It Through Real Tests
First test: I gave it my entire codebase for a side project (about 50 files, 5,000 lines of code) and asked it to identify potential bugs. It found three legitimate issues I'd missed, including one that would've caused problems in production. Not bad.
Second test: Multilingual support. I work with a team that speaks English, Spanish, and Mandarin. Fed it meeting transcripts in all three languages mixed together. It summarized the key points accurately and even caught nuances that pure translation wouldn't capture.
Third test: The real-world one. I'm planning a trip to Japan and needed to compare hotels, check reviews, figure out transportation options, and build an itinerary. With the Deep Research feature now on mobile, I could do all of this while commuting. It generated a comprehensive report in minutes.
Someone I know at a content studio tried using it for video analysis. They've been manually tagging and categorizing thousands of video clips. Gemini 2.0 churned through them and created a searchable database in a fraction of the time it would've taken their team.
The Free Tier Is Surprisingly Good
Here's what shocked me: the free version isn't gimped. You get access to 2.0 Flash with the 1 million token context window. You can upload files, images, videos. You get the newer, faster model that honestly competes with ChatGPT Plus in many ways.
Obviously there are usage limits—you can't hammer it with thousands of requests per day for free. But for normal use? It's generous. Google clearly wants adoption more than they want subscription revenue right now.
I've been using ChatGPT Plus for a year. I'm not canceling it yet, but I'm definitely using Gemini more than I expected to. For certain tasks—especially anything involving large documents or multimodal inputs—it's actually better.
Where It Still Falls Short
The conversational flow isn't as natural as ChatGPT. It feels more like talking to a very smart search engine than having a dialogue. For creative writing or brainstorming, I still prefer ChatGPT.
The mobile experience, while improved, still lags behind ChatGPT's app. The Android integration is neat if you're in the Google ecosystem, but the standalone app feels clunkier.
And I'm still seeing occasional hallucinations, though fewer than before. Asked it about some niche technical details in my field and it confidently stated something completely wrong. Always verify, folks.
The Competitive Picture
This move puts real pressure on OpenAI. If Google's giving away a model that's competitive with ChatGPT Plus for free, why would casual users pay $20/month? OpenAI's advantages are the interface, the ecosystem, and features like DALL-E and Advanced Voice.
But Google has advantages too. Integration with Workspace, the massive context window, and honestly, better handling of complex documents and media. Plus they're burning money to gain market share, and they've got deeper pockets than almost anyone.
Anthropic's Claude is still my favorite for certain tasks—it's more careful, better at reasoning, and the interface is cleaner. But Claude's free tier is limited and Pro costs the same as ChatGPT Plus. Gemini 2.0 being free changes the calculation.
The Catch Nobody's Mentioning
Here's the thing: Google isn't doing this out of the goodness of their hearts. Free access means they're collecting data, training models on usage patterns, and building profiles. It's the same model as Gmail and Search—give away the product, monetize the data.
I'm okay with that tradeoff personally, but it's worth being aware of. If you're using Gemini for sensitive work stuff, maybe spring for the paid tier which has better privacy protections. Or use Claude or ChatGPT where you're paying for the service instead of being the product.
Should You Switch?
If you're not paying for AI right now and just using the free ChatGPT tier? Absolutely try Gemini 2.0. It's better in most ways.
If you're paying for ChatGPT Plus? Maybe not worth switching completely, but definitely worth having both. Use Gemini for research and document analysis, ChatGPT for conversation and creative work.
If you're all-in on Claude? Honestly, probably stay there. Claude's still the most reliable for serious work where accuracy matters.
What This Means For the Industry
Google going aggressive with free access is fascinating. They're essentially subsidizing AI usage to win market share. That's good for consumers—we get better tools for free—but it might squeeze out smaller players who can't compete on price.
It also signals that the AI race is far from over. Every time someone looks dominant (OpenAI with GPT-4, then DeepSeek with cost efficiency), someone else makes a move that shifts the landscape. Google was looking like they might fall behind permanently. This puts them back in the game.
My Take
I'm genuinely excited about this in a way I haven't been about Google AI in a long time. Gemini 2.0 feels like they finally figured out what they're doing. It's not perfect—nothing is—but it's legitimately competitive and in some areas better than the alternatives.
The free access changes everything. I've already recommended it to three friends who were on the fence about trying AI tools. When there's no barrier to entry and the quality is this good, why wouldn't you at least experiment?
Will I still use multiple AI tools depending on the task? Yeah, probably. But Gemini 2.0 is now in my regular rotation instead of being the backup option I almost never used.
Google might've just pulled off one of the biggest comeback stories in tech. Or at least in AI. Either way, they've got my attention again.