ChatGPT vs Claude vs Gemini: Which AI Writes the Best CV?

Tailoring your CV to every job listing used to be a hassle. With AI, it's easier than ever — but which AI actually does it best? I gave ChatGPT, Claude, and Gemini the exact same CV and job listing to rewrite. The results surprised me — especially when one of them started making up skills the applicant never had.

The Experiment

How well can ChatGPT, Claude, and Gemini rewrite a CV?

I found a UX designer job listing on Indeed and drafted a fairly bare-bones CV for a fictional applicant. The CV was intentionally weak:

• The applicant's experience didn't fully meet the requirements (to see if the AIs would make things up)

• Vague job descriptions

• Casual wording

• Messy structure

I ran all three AIs on the same prompt simultaneously for a direct comparison.

(Note: ChatGPT 5.2, Claude Sonnet 4.6, Gemini 3 Pro)

The Prompt

"I'm applying for this job. Can you review my CV and tell me how I should improve it? Job listing: [pasted here]"

Scoring Criteria

I evaluated each AI's output on four dimensions, each scored out of 5:

1.    Honesty & Tailoring — does it tailor the CV to the job while staying true to the applicant's actual abilities?

2.    Structure & Organization — is the CV well-organized and easy to scan?

3.    Writing Quality — is the language professional, concise, and easy to read?

4.    Human-sounding — does it sound like a real person wrote it, or does it scream AI?

The Results

1. Honesty & Tailoring

Gemini — 4/5: Best balance. Stayed mostly true to the applicant's skills while tailoring to the job. It did make some reasonable assumptions that weren't explicitly stated.

ChatGPT — 3/5: Very honest, but barely tailored. Didn't add many keywords from the posting. Safe, but a missed opportunity — a hiring manager wouldn't immediately see why the applicant fits this role.

Claude — 1/5: Most aggressively tailored, but largely fabricated. Nearly every keyword from the job listing appears, but many were made up. It essentially rewrote the applicant into a different person.

2. Structure & Organization

ChatGPT — 5/5: Cleanest structure. Consistent formatting and logical ordering. Easy to scan.

Gemini — 4.5/5: Also reorganized well. Condensed the CV into a tight, scannable format. Minor formatting inconsistencies, but overall strong.

Claude — 3.5/5: Mostly followed the original order, which isn't ideal. Added a soft skills section (a nice touch), but some structure choices were confusing and disrupted the flow.

3. Writing Quality

Gemini — 5/5: Professional and easy to read. Highlighted relevant keywords without overdoing it, making the fit clear.

Claude — 4/5: Strong writing with nice details — if they were actually true. But sometimes added unnecessary specifics that made it less concise.

ChatGPT — 3/5: Stayed closest to the original wording. Didn't really improve tone or flow. Reads more like an edited draft than a polished CV.

4. Does It Sound Human-Written?

ChatGPT — 5/5: Closest to the original voice. Reads like the applicant actually wrote it.

Gemini — 4.5/5: Natural and polished, though the summary felt slightly generic.

Claude — 4/5: Professional, but sometimes overly polished in a way that might raise a flag.

(Top row: ChatGPT; Bottom left: Claude; Bottom right: Gemini)

Final Scores

1.  Gemini: 18/20

2.  ChatGPT: 16/20

3.  Claude: 12.5/20

Verdict

Gemini wins on overall balance. It produces a CV that feels both tailored and professional, while still sounding natural.

ChatGPT is the safest option. It stays true to your original content, but doesn't go far enough to really elevate it.

Claude produced the most impressive-looking CV — but also the least trustworthy. It included specific metrics and skills that don't exist. If those were real, it would be an excellent CV.

Practical Tip: How to Prompt for a Better CV

If you're using AI to rewrite your CV, try this prompt:

"Rewrite my CV for this job. Only include skills I actually have, but reframe them to fit the role. Make it professional, skimmable, and optimized with keywords from the job listing. Keep my original voice — don't over-polish."

 The truth? Each AI has its own strengths and weaknesses, as you just saw. The best answer rarely comes from just one. That's actually why I built Agora. It lets you ask all three at once, then carefully analyzes their responses and gives you one clear answer.

What's Next?

Next up: I'm testing all three AIs on cover letters. Follow The Agora Lab for more comparisons.