Closed diderich closed 2 years ago
Hi @diderich, good question, thanks. We're actively working on improving the speed of our service. I hope I can answer some of your questions:
We prioritize Pro requests higher than Free requests, so you may see a performance improvement by upgrading. However the speed depends on many other factors including the demand on our services at any moment.
Yes, some of the speed difference may be explained by location: while Google Cloud has regions in Zurich and Frankfurt, our servers are located in Finland for data security reasons.
I am currently benchmarking DeepL against Google Translate for translating photo captions the size of a Tweet with a rather standardized sentence structure from English to German and French. 80% of the translated caption phrases are identical and in the last 20%, DeepL chooses a more appropriate vocabulary, e.g.
But I noticed that on the speed/latency side, Google is 2x to 3x faster (0.15s vs 0.31s, 0.13s vs 0.39s, 0.2s vs 0.52s, etc.). I was wondering if that is due to