We have an API Service that is utilizing Perplexity’s chat completion via the official NodeJS SDK. We’re using the sonar-pro model with high reasoning effort. With no changes on our end, we’ve started noticing a huge increase in latency from perplexity API, we were originally averaging requests in the 3-10 second range and now we’re seeing 30+ second requests with the exact same prompt / configurations. Has something changed internally and is this is expected? is there a service degradation that we should be aware of? I am happy to provide any additional details or information to help figure this out, just hoping to get some clarity as this is impacting our customers.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| API very slow / returning 5xx errors | 2 | 145 | May 15, 2025 | |
| Api request speed is very slow, is there any solution？ | 2 | 411 | May 15, 2025 | |
| Perplexity charged $5 (USD) in only 03 requests | 0 | 34 | November 10, 2025 | |
| After updating: Network timeout at: https://api.perplexity.ai/chat/completions | 4 | 261 | May 15, 2025 | |
| Subpar responses when using pplx api when compared to perplexity labs | 2 | 136 | May 15, 2025 |