A compact text-to-speech model in the GPT-4o family optimized for low-latency audio generation.
No resource links recorded.