Whisper Large v3 Turbo (speech-to-text)

Whisper Large v3 Turbo debuts with strong ASR performance baseline

Whisper Large v3 Turbo by OVH AI Endpoints has entered benchmarking with its initial performance baseline for automatic speech recognition. As a newly tracked model, this represents the first measurement window, establishing reference metrics for future comparison. The model is based on OpenAI's Whisper Large v3 architecture optimized for faster inference while maintaining transcription quality. This turbo variant is designed to balance accuracy with reduced latency compared to the standard Large v3 model. Since this is the debut benchmark window, there are no performance trends or changes to report yet. Users should consider this a starting point for evaluating the model's ASR capabilities. The service is deployed through OVH's GRA region infrastructure, providing European-based hosting for speech-to-text workloads. Future benchmark windows will reveal performance stability, any quality improvements or regressions, and how the model compares against competing ASR solutions. Organizations evaluating this model should monitor subsequent benchmarks to understand its consistency and reliability over time before committing to production deployments requiring predictable speech recognition performance.