llm-benchmarks
githubdrose.io

perplexity Provider Benchmarks


Measured over the last 30 days

Models Tracked
4
Avg Tokens / Second
20.46
Avg Time to First Token (ms)
5022.50
Last Updated
Oct 6, 2026

All perplexity models

4 live · one row per model · fastest first
perplexity site
ModelAvg Toks/SecMinMaxAvg TTF (ms)
Sonar38.7038.7038.701270.00
Sonar Pro25.4025.4025.402130.00
Sonar Pro Search13.1013.1013.104190.00
Sonar Reasoning Pro4.654.654.6512500.00

Frequently Asked Questions

Which perplexity model is fastest?
Based on recent tests, Sonar shows the highest average throughput among tracked perplexity models.
How many recent measurements feed this dashboard?
This provider summary aggregates 4 individual prompts measured across 3 monitoring runs over the past month.