GPT-4o mini TTSvsQwen3-TTS
Speed and accuracy
Left is faster. Higher reads more hard parts right.
- GPT-4o mini TTS688 ms · 82.7%
- Qwen3-TTS1.65 s · 71.0%
GPT-4o mini TTS is better on non-streaming latency, accuracy and naturalness. Qwen3-TTS is better on price.
Last updated Methodology by Dileep Chagam
Which is better at what
Measure by measure, the same sentences, the same place.
MeasureGPT-4o mini TTSQwen3-TTSBetter
- Non-streaming latency (median)GPT-4o mini TTS is better688 ms1.65 sGPT-4o mini TTS
- Streaming latency (median)Can't streamCan't streamn/a
- Extra audio after a stopNo stop messageNo stop messagen/a
- Hard parts read rightGPT-4o mini TTS is better82.7%71.0%GPT-4o mini TTS
- Naturalness (1 to 5)GPT-4o mini TTS is better4.434.34GPT-4o mini TTS
- Price per 1M charactersQwen3-TTS is better$25.70$20.00Qwen3-TTS
Hard parts read rightGPT-4o mini TTSQwen3-TTSBetter
- Numbers and shorthand7 a.m. · 20% · 36°F · approx.78%69%GPT-4o mini TTS
- Codes and addressesSAVE20 · help@north-clinic.org · 1-800-555-019978%49%GPT-4o mini TTS
- Read from contextSt. Mary's on Main St. · Dr. Okafor96%75%GPT-4o mini TTS
- Medical termslevothyroxine · gastroenterology94%100%Qwen3-TTS
An estimate: speech-to-text transcribers scored it, not people.
Hear both
The same sentence, read by each voice.
19 sentences, the same for bothGPT-4o mini TTSQwen3-TTS
- 01A greetingThanks for calling Cedar Valley Family Practice, this is Riley. How can I help you today?
- 02MoneyYour current balance is $1,347.09.
- 03A dateThe order was placed on 2026-03-14 and shipped two days later.
- 04Tracking codeThe tracking number is 1Z 7A4 X29 0341 8867.
- 05Email addressI have m.rivera88@example.org, is that still current?
- 06An Irish nameSiobhan Ó Briain called about the invoice.
How each ranks among all 26
MeasureGPT-4o mini TTSQwen3-TTS
- Non-streaming latency (median)21 of 2626 of 26
- Streaming latency (median)n/an/a
- Extra audio after a stopn/an/a
- Hard parts read right12 of 2625 of 26
- Naturalness (1 to 5)7 of 2619 of 26
- Price per 1M characters11 of 258 of 25
GPT-4o mini TTS vs Qwen3-TTS FAQ
Which starts talking sooner, GPT-4o mini TTS or Qwen3-TTS?
GPT-4o mini TTS. GPT-4o mini TTS: 688 ms. Qwen3-TTS: 1.65 s.
GPT-4o mini TTS. GPT-4o mini TTS: 688 ms. Qwen3-TTS: 1.65 s.
Neither has a figure. GPT-4o mini TTS: Can't stream. Qwen3-TTS: Can't stream.
Neither has a figure. GPT-4o mini TTS: No stop message. Qwen3-TTS: No stop message.
GPT-4o mini TTS. GPT-4o mini TTS: 82.7%. Qwen3-TTS: 71.0%.
Qwen3-TTS. GPT-4o mini TTS: $25.70. Qwen3-TTS: $20.00.