Can you recommend a good serverless GPU provider that supports running WhisperX?

[yccheok]

Here are my test results so far. None have been successful yet:

RunPod – Satisfied with their faster-whisper pre-built template in terms of service quality and cost. However, I’m facing issues building https://github.com/yccheok/whisperx-worker on their serverless solution. Still waiting for a response from customer support.

Beam Cloud – Way more easier to setup than RunPod. Unsatisfied with the service quality. A significant percentage of tasks remain stuck in the "pending" state indefinitely. Also, the pricing lacks transparency, showing costs 10× higher than expected.

Fireworks – No setup required. Unsatisfied with the service quality. (Tested with OpenAI Whisper Turbo V3, not WhisperX.) The service went down several times during testing, and support records show this happens multiple times per month.

If you have experience running WhisperX in a serverless environment, can you recommend a reliable service provider?

Thank you.