NVIDIA Nemotron 3.5 Transcribe vs GPT-4o Transcribe
Accuracy and speed
Left gets fewer words wrong. Lower has the final text ready sooner. Nemotron 3.5 Transcribe doesn't support streaming, so it is not shown.
- Nemotron 3.5 Transcribe1.89%
- 4o Transcribe3.90% · 652ms
NVIDIA Nemotron 3.5 Transcribe is ahead of GPT-4o Transcribe on both datasets together, the Pipecat Dataset and the Ocular Dataset.
Last updated Methodology by Dileep Chagam
Who is better at what
- Overall WER2.28%6.85%
- Pipecat Dataset WER1.89%3.90%
- Ocular Dataset WER3.03%12.45%
- Time to first textn/an/a
- Final-text delayn/a652ms
Lower is better on every measure. Too close to call means the likely ranges overlap.
Nemotron 3.5 Transcribe vs 4o Transcribe FAQ
Which is more accurate, NVIDIA Nemotron 3.5 Transcribe or GPT-4o Transcribe?
NVIDIA Nemotron 3.5 Transcribe is more accurate on both datasets. On the 1,000 public Pipecat clips, NVIDIA Nemotron 3.5 Transcribe scores 2.28% and GPT-4o Transcribe 6.85%, so NVIDIA Nemotron 3.5 Transcribe makes fewer errors by 4.57 points. On the eight licensed Ocular recordings, NVIDIA Nemotron 3.5 Transcribe scores 1.89% and GPT-4o Transcribe 3.90%, so NVIDIA Nemotron 3.5 Transcribe makes fewer errors by 2.02 points. Word error rate counts incorrect, missing, and extra words, so lower is better.
NVIDIA Nemotron 3.5 Transcribe is more accurate on both datasets. On the 1,000 public Pipecat clips, NVIDIA Nemotron 3.5 Transcribe scores 2.28% and GPT-4o Transcribe 6.85%, so NVIDIA Nemotron 3.5 Transcribe makes fewer errors by 4.57 points. On the eight licensed Ocular recordings, NVIDIA Nemotron 3.5 Transcribe scores 1.89% and GPT-4o Transcribe 3.90%, so NVIDIA Nemotron 3.5 Transcribe makes fewer errors by 2.02 points. Word error rate counts incorrect, missing, and extra words, so lower is better.
They cannot be compared on speed. NVIDIA Nemotron 3.5 Transcribe was tested on whole recordings rather than a live stream, so its speed is not measured.
NVIDIA Nemotron 3.5 Transcribe makes fewer errors on real conversations. NVIDIA Nemotron 3.5 Transcribe was tested on whole recordings rather than a live stream, so its speed is not measured. A voice agent cannot reply until the final transcript arrives, and the Ocular recordings are real conversations.
NVIDIA Nemotron 3.5 Transcribe was tested on whole recordings rather than a live stream, so whether it streams partial text was not measured. GPT-4o Transcribe returns no text until the speaker stops.
Test your voice agent the same way.
Cekura runs these simulated calls on your agent and scores every one.