My Honest Take After 3 Months of Using AI Tools

It Started with ChatGPT

I started using AI tools seriously last November. My original plan was to just use ChatGPT — but then people kept telling me Claude was great for writing, Gemini leveraged Google's search strength, and Copilot was free. I figured, why not just try them all?

So for three months I rotated between ChatGPT, Claude, Gemini, and Copilot. I paid for subscriptions on some, used free tiers on others. Here's my honest, unfiltered experience.

ChatGPT — Still the Benchmark

Most-used tool: ChatGPT (GPT-4o)

Honestly, ChatGPT felt the most natural. It just flows. Here's what I used it for most:

• Coding questions: "Why is this error happening?" — instant answer. I asked about Python code constantly and found the accuracy around 95%.
• Blog drafts: Give it keywords and it structures the post for you. I always rewrite in my own style, but having a scaffold is useful.
• Brainstorming: Ask "what do you think about building this kind of service?" and it actually engages with the problem.
• Translation: When I hit a wall in an English document, it understands context in a way Google Translate simply doesn't.

Pros:

• Conversation feels natural, almost like talking to a person
• Strong creative output
• Plugins cover web search and image generation
• Fast update cadence

Cons:

• Hallucinations happen. Be especially careful with recent facts
• GPT-4o requires the paid plan ($20/month)
• English-centric; can feel slightly off in other languages sometimes

Claude — A Writing Powerhouse

I tested Claude Opus 4.5 and was genuinely surprised. I fed it a 20-page PDF report and asked for a summary — it extracted the core points in under 5 minutes. ChatGPT required splitting the file due to size limits; Claude handled it in one pass.

Primary use cases:

• Long document analysis: Reads 200-page papers. Not an exaggeration.
• Writing quality: More polished prose than ChatGPT, with fewer grammatical slip-ups
• Code review: Pinpoints problems in my code with precision
• Nuanced reasoning: Ask "what's the right approach in this situation?" and you get multiple well-considered perspectives

Pros:

• Long-context processing is exceptional (up to 200K tokens)
• Best-in-class writing quality
• Low hallucination rate
• Natural, well-calibrated tone

Cons:

• Slower than ChatGPT; responses take noticeably longer
• Slightly less creative flair compared to GPT
• No image generation
• Also $20/month

Gemini — Google's Depth

Gemini stood out most during research. It's Google — search is in its DNA. Ask "summarize 2026 AI trends" and it pulls recent articles and cites them inline. Real-time web access by default.

Primary use cases:

• Current information: Real-time queries it handles natively
• Google Workspace integration: Seamless with Gmail and Google Docs
• YouTube summarization: Paste a YouTube link and get a content summary
• Multimodal tasks: Handles images and text simultaneously

Pros:

• 1 million token context window (handles extremely long documents)
• Deep integration with the Google ecosystem
• Web search included by default
• Free tier is genuinely usable

Cons:

• Conversational flow can feel choppy at times
• Coding assistance lags behind ChatGPT and Claude
• If you're privacy-conscious, having your data go through Google gives some people pause

Copilot — The Free Option

I didn't expect much from Microsoft Copilot, but it surprised me — mostly because it's free. Available natively in Edge and Windows, no setup required.

Primary use cases:

• Quick lookups: Simple questions handled fast
• Microsoft 365 integration: Works directly in Word, Excel, and PowerPoint
• Enterprise document work: Useful for structured business data

Pros:

• Free
• Optimized for Microsoft ecosystem users
• Fast response times

Cons:

• Shallow on complex questions
• GPT-based but feels limited
• Limited creative output

3-Month Comparison Table

Where I Was Disappointed

Being honest about the frustrations:

1. They all say similar things I ran the same questions through all four tools and the answers were often nearly identical. For general knowledge questions it almost felt copy-pasted. "They're all the same underneath" crossed my mind more than once.

2. Hallucinations remain a real issue I asked ChatGPT about Olympic medal standings and got completely wrong information. Always fact-check time-sensitive or specific claims.

3. Creativity has a ceiling Ask for "a genuinely novel idea" and you'll get variations on familiar themes. Truly original thinking still comes from people.

Where I Was Genuinely Impressed

The upsides were real:

1. Coding speed tripled I'm not exaggerating. Things that took an hour of Stack Overflow digging now take 10 minutes. Paste an error message, get a fix. Development velocity is measurably higher.

2. Language barriers dissolved English papers, technical documentation in other languages — paste it to ChatGPT and get a context-aware explanation, not just a word-for-word translation.

3. A 24/7 mentor Debugging at 3am and hitting a wall? Someone to ask. Questions you'd feel embarrassed asking a colleague? Ask freely, no judgment.

4. Structured thinking Dump half-formed ideas at ChatGPT and ask it to "organize this into a structure" — it builds a coherent framework you can actually work from.

Conclusion: My Current Workflow

After three months, here's how my usage actually breaks down:

• Daily questions, coding: ChatGPT (70% — highest versatility)
• Long documents, writing: Claude (20%)
• Current information: Gemini (5%)
• Quick lookups: Copilot (5%)

ChatGPT became my default because of its versatility — it handles most things well enough. But the key insight is: using only one tool is leaving value on the table. Each has different strengths, and matching the tool to the task is the real unlock.

If you're still deciding where to start:

• Budget-conscious → Start with Copilot (free)
• Coding and general use → ChatGPT
• Writing and analysis → Claude
• Research and search → Gemini

What AI Tools Are You Using?

That's my three-month experience. What's your go-to AI tool, and for what? If you have a tip like "this tool is best for X specific use case," I'd love to hear it in the comments. Next up: a deeper head-to-head comparison of ChatGPT vs Claude. Thanks for reading!