Gradium vs GPT-4o Mini Transcribe
Accuracy and speed
Left gets fewer words wrong. Lower has the final text ready sooner.
- Gradium6.48% · 212ms
- 4o Mini Transcribe3.35% · 630ms
GPT-4o Mini Transcribe is ahead on both datasets together and the Pipecat Dataset, and Gradium on final-text delay. The two tie on the Ocular Dataset.
Last updated Methodology by Dileep Chagam
Who is better at what
- Overall WER5.78%3.73%
- Pipecat Dataset WER6.48%3.35%
- Ocular Dataset WER4.45%4.45%
- Time to first text1.71sn/a
- Final-text delay212ms630ms
Lower is better on every measure. Too close to call means the likely ranges overlap. Our guide explains how word error rate is measured.
Gradium vs 4o Mini Transcribe FAQ
Which is more accurate, Gradium or GPT-4o Mini Transcribe?
GPT-4o Mini Transcribe is more accurate on both datasets. On the 1,000 public Pipecat clips, Gradium scores 5.78% and GPT-4o Mini Transcribe 3.73%, so GPT-4o Mini Transcribe makes fewer errors by 2.05 points. On the eight licensed Ocular recordings, Gradium scores 6.48% and GPT-4o Mini Transcribe 3.35%, so GPT-4o Mini Transcribe makes fewer errors by 3.12 points. Word error rate counts incorrect, missing, and extra words, so lower is better.
GPT-4o Mini Transcribe is more accurate on both datasets. On the 1,000 public Pipecat clips, Gradium scores 5.78% and GPT-4o Mini Transcribe 3.73%, so GPT-4o Mini Transcribe makes fewer errors by 2.05 points. On the eight licensed Ocular recordings, Gradium scores 6.48% and GPT-4o Mini Transcribe 3.35%, so GPT-4o Mini Transcribe makes fewer errors by 3.12 points. Word error rate counts incorrect, missing, and extra words, so lower is better.
Final-text delay was not measured for one or both, so speed cannot be compared. Both are measured on the same 206 streamed turns.
GPT-4o Mini Transcribe makes fewer errors on real conversations, and the two are not separated on speed. A voice agent cannot reply until the final transcript arrives, and the Ocular recordings are real conversations.
Only Gradium. GPT-4o Mini Transcribe returns no text until the speaker stops, which matters if your agent reacts to partial transcripts.
Test your voice agent the same way.
Cekura runs these simulated calls on your agent and scores every one.