At this point, everybody has one favourite AI app they use for work.

I have taken it a bit further. This year, I assembled (no pun intended) my own AI agent dream team, a group featuring seven of the best AI models around. Each has its own special characteristic, personality and speciality.

A lot of the time, I even use the same prompt in all of them just to see what different answers they show me. It’s like a group of experts debating ideas, in real-time, everyone with an opinion and a style.

And because all of these models are at or near the top in LLM leaderboards, I know that it’s giving me some of the best reasoning, knowledge and insight out there at the moment. Below is the full lineup of my team of AI agents, how I utilise them, what they excel at and why you might want to build some yourself.

1. ChatGPT The All-Around Powerhouse

ChatGPT is my all-purpose assistant. It’s excellent for the coders, thinkers and content writers, planning daily activities or creating images. Whether I’m writing code or brainstorming ideas, ChatGPT gets the job done.

The recent models also have more of a talkative personality, reason better and can handle longer contexts. I use it as my daily driver for most tasks like productivity, general tasks and coding. On mobile, it’s my favourite for fast research and idea dumps when I’m on the go. It’s the way your reliable co-worker always comes through because it’s fast, smart and well thought out. It’s also very good at generating images.

2. Claude The Calm Strategist

Claude can act as a second brain. It’s great for structured reasoning, long-form writing and big-picture thinking. Even though ChatGPT is great for problem-solving and quick tasks, Claude sometimes shines when I need to take a step back and dig deeper, like when analysing docs, decoding complex concepts or brainstorming projects.

It’s also still one of the best coding LLMs out right now, especially when it comes to producing clean and readable code. If I’m focusing on an article, a tech review or a deep work session, Claude is a good choice.

3. Qwen The Researcher

Qwen is one of the up-and-coming stars in my AI rotation. It’s a great tool for data collection, analysis and research. It can hunt down citations, detail studies and sort through niche technical topics pretty well.

It’s also good at coding, maybe not as slick as ChatGPT or Claude, but fast and flexible. I use Qwen when I want an alternative or want to see a different perspective on something.

ChatGPT and Claude often take the lead; Qwen is a great alternative for some tasks or when the others are busy.

4. Grok The Real-Time News Expert

Grok, developed by xAI, reads directly from Twitter/X. Basically, for me, it’s the AI news engine that helps keep me up to date on tech, AI and world events, all in real time.

I use it to:

• Track trending topics
• Receive instant inspiration while browsing
• Summarise what’s happening in tech, news, and everything else

It’s also great for live research and discovery because it is plugged into the conversations going on online. If I want to know what’s hot or what the devs are talking about today, Grok is the one.

5. Gemini The Polished Problem-Solver

Gemini seems like a natural multitasker. I find that it is a great tool for structured problem-solving, summarisation and quick fact questions.

It’s not always as wild and inventive as ChatGPT or Claude, but it’s solidly reliable, and of course, so deep in Google’s ecosystem, which makes it ideal for when I’m researching notes in Google Docs or Drive, or even browsing YouTube videos.

I use Gemini all the time when I’m looking for a second opinion on something technical or factual. I guess you could say it’s my “grounded” agent, which is less opinionated and more balanced. Image creation is right up there with the best platforms, too.

6. DeepSeek The Document Analyst

DeepSeek is one of my favourite tools for when you need something more like an ask-heavy research task. It’s pretty impressive because you can upload and then analyse large numbers of files.

I use it when I’m doing document insight studies or comparing any large reports or technical papers. When I need to sort through a complicated dataset or a research paper, DeepSeek is my data scientist because it can handle multiple attachments more so than other LLMs and all for free.

7. Perplexity The AI Search Engine

I turn to the A.I., based search engine Perplexity for web searches all the time these days. It’s like Google on steroids because it goes deeper, is sourced, and explains things well.

I use it when I need to look something up quickly or verify facts. It’s great for brainstorming too, as I can search for something like “best JavaScript frameworks”, and it will condense everything and even create a summary in a table.

It’s less like talking to a chatbot and more like talking to a research assistant who can think critically. Perplexity is also baked into the Comet Browser, so I can use it natively in my primary browser all the time.

How I Use Them Together

Here’s the fun part: I don’t use these tools separately; I treat them like an AI team.

When I’m exploring a topic, I will most times use the same prompt in multiple model tools like for example, asking all of them:

“What are the key trends shaping AI development in 2025?”

Each one responds differently, like they are individuals in a team:

• ChatGPT might give a broad, structured overview
• Claude offers a calm, reasoned breakdown
• Qwen pulls in research-heavy details
• Grok gives me the latest trending context
• Gemini adds factual precision
• DeepSeek analyses data sources
• Perplexity wraps it all up with verified references

It’s like having a debate between world-class experts, and I get to listen to all sides before forming my own conclusion. I even have the same setup on my phone because I have the mobile apps for all of them, so I can repeat the same process on the go. It truly feels like having a team of AI agents or team members everywhere you go.

Conclusion

This year, a single AI model is no longer enough. We have a huge choice of LLMs and each has its own strengths, personality and bias, and that’s why I use many of them together. Viewing them as a panel of experts makes for fast, easy and deep learning and information gathering.

And because all of these models are trying to be at the top of the global LLM leaderboards, I know I’m getting best-in-class reasoning and the most up-to-date insights.

AI isn’t changing the way we work; it’s improving the way we think. And you know what that really and truly is, a lot of people’s most productive upgrade in many years.