FOWL AI Tests · Monthly · Test #3

We asked three AI assistants for a real voiceover file. Two delivered one.

ChatGPT, Claude, and Gemini: same fictional app launch, same prompt, word for word, asking for a 60-second script and the actual audio file. We downloaded every file, measured it, and transcribed it ourselves. Here's what actually showed up.

Tools testedChatGPT (Pro, default), Claude (Opus 5.5, default), Gemini (free tier, Flash)
Test case60-second launch voiceover for a fictional chores app
What we checkedThe real audio file: length, pauses, loudness, transcript
Key Findings — 15-Second Version
  • ChatGPT hit the brief almost exactly: a 142-word script and an audio file that runs 60.46 seconds, in both MP3 and WAV.
  • Claude also delivered a real MP3 (57.5 seconds) plus a script file, and wrote the funniest script of the three.
  • Gemini wrote a script but said it can't produce audio. It also called its script 148 words when it's 127, below the 140–160 we asked for.
  • All three covered every point in the brief and invented no statistics, reviews, or features.

We invented a fictional app, Choreboard, a free iPhone and Android app that rotates household chores and asks for a photo when one is done, and gave each assistant the exact same brief. The ask had two parts: write a 60-second explainer script, and produce the actual audio file of it being read aloud, as a download. Each tool ran on its default model in a normal consumer account. No follow-up prompts, no edits: we took the first thing each tool produced.

Then we checked the files, not the chat replies: duration and format with ffprobe, pauses and peak loudness with ffmpeg, and a transcript from Whisper to confirm the audio actually says the script, including the product name.

ProductFictional, invented for this test
PromptIdentical across all three, word for word
Output checkedThe downloaded audio file, not the chat reply
VerificationMeasured duration, pauses, loudness; transcribed
Show the exact prompt we used
I need a 60-second product-explainer voiceover for a fictional app launch. Please (1) write the script and (2) produce the actual audio file of it being read aloud, as a downloadable file. The brief: - App name: Choreboard - What it is: a free iPhone and Android app for roommates and families to split household chores fairly. - How it works: everyone in the home joins one shared board; chores rotate automatically each week; when someone finishes a chore they tap it and snap a quick photo, and everyone else sees it's done. - Who it's for: people tired of arguing about whose turn it is to take out the trash. - Launch date: October 6, 2026. - Call to action: "Download Choreboard free on the App Store and Google Play." - Tone: warm, a little funny, not salesy. Requirements: the finished audio should run close to 60 seconds (roughly 140 to 160 words). Do not invent statistics, user counts, reviews, awards or features that aren't in the brief. Give me the script as text and the audio as a file I can download.

Fine print. We planned to include ElevenLabs, but it only turns a finished script into speech; it doesn't write one from a brief, so it isn't a like-for-like comparison. Gemini ran on the free tier (Flash) because our paid plan had lapsed; a paid Gemini plan may behave differently. Claude first asked to use a third-party voice service we had connected to the account; we declined, so the result reflects Claude on its own.

Five categories, scored from the actual files.

Every category is scored 1–5, based only on what we found in the real output, never the chat reply.

5 = Excellent 4 = Good 3 = Acceptable 2 = Weak 1 = Poor
CategoryWhat it measures
Real DeliverableDid it hand over a downloadable audio file and the script, as asked?
TimingHow close the audio runs to 60 seconds, and whether the script lands in the 140–160 words we asked for.
Delivery & PacingMeasured from the file: speaking rate, pauses, clipping, and whether the transcript matches the script.
Script CraftWarm, a little funny, not salesy, as the brief asked, and still covers every point in it.
HonestyNo invented facts, and no wrong claims about its own output (word count, length, what it could do).
Category
ChatGPT
Claude
Gemini
Real Deliverable
Timing
Delivery & Pacing
Script Craft
Honesty
Total (out of 25) 24
Winner
23 10

Same brief, three different opening lines.

The first line of each script, verbatim, no edits.

ChatGPT

"Household chores shouldn't require a group chat investigation."

Claude

"Quick question: whose turn is it to take out the trash?"

Gemini

"Let's be honest: nobody woke up today hoping to have another debate about whose turn it is to take out the trash."

The actual audio files. Press play yourself.

"Give me the audio file" is now a real ask, for two of the three
ChatGPT and Claude both produced a playable MP3 in one pass, with no follow-up prompt and no outside service. Gemini, on the free tier, stopped at the script. If audio is the deliverable, that difference matters more than any line of copy.
The self-reported numbers were right twice, and wrong once
ChatGPT said 142 words and about 60.5 seconds: we counted 142 words and measured 60.46 seconds. Claude said 156 words and 57.5 seconds: we counted 157 and measured 57.49. Gemini said its script was 148 words "timed to run right around 60 seconds." We counted 127, below the range we asked for; read at the ~150 words a minute Gemini itself assumed, that's closer to 51 seconds. If a tool tells you its script is the right length, count it.
Nobody padded the brief
We tested this on purpose: the brief contained no user counts, ratings, or awards, and the prompt said not to invent any. All three scripts covered every point we checked (the name, free, iPhone and Android, the shared board, weekly rotation, tap-and-photo, the launch date, the exact call to action) and none added a statistic or a feature. The only drift: Gemini widened the audience to "partners," which the brief didn't mention.
🏆 The Verdict
Best overall
ChatGPT

Right length, right pace, two file formats, and every claim about its own output was accurate. The one to use when you need a usable file on the first try.

Best script
Claude

The warmest, funniest writing of the three, with a real MP3 and a separate script file. Its pacing is quicker between pauses, so you may want to ask it to slow down.

The gap between ChatGPT and Claude is one point and mostly about timing. The real gap is whether a tool can produce audio at all. We didn't compare against dedicated voice tools or a human voice actor this time, so treat both as fast first drafts you can hear today, not a final read.

Should you switch tools?
Use ChatGPT if you want a ready-to-use audio draft that hits a set length.
Use Claude if the script matters most and you're happy to tweak the pace.
Use Gemini for the script and timing cues, then take it to a text-to-speech tool yourself.

We test a new AI tool matchup every month.

Real files, real numbers, no sponsorships. Subscribe to get the next one first.

Subscribe free →