In short
Whisper v4 is a state-of-the-art, open-source automatic speech recognition model by OpenAI. It uses a transformer architecture to transcribe and translate speech in over 100 languages with high robustness to noise.
Frequently asked
- What is the primary function of Whisper v4?
- It is an automatic speech recognition model that converts audio into text. It also performs speech translation into English without requiring separate language identification.
- How many languages does Whisper v4 support?
- The model supports over 100 languages using a single architecture. It is noted for strong performance in low-resource languages.
- Is Whisper v4 available for free?
- Yes, it is open-source and freely available. An API option is also listed at $0.006 per minute.
- What are the main limitations of Whisper v4?
- It can be computationally intensive, lacks native speaker diarization, and does not directly support streaming audio input. Performance may degrade in extremely noisy or highly accented speech.