Performance, in context.
The useful measure is how long you wait for an answer during a conversation.
From question to first word
We measure the time from the end of a spoken question to the first word appearing in the overlay, and the time until the full answer is ready. Both include transcription.
Know the setup behind the result
Each result will identify the PC, GPU, CueFlect edition, answering mode and models. That context helps you judge whether a result is relevant to your own setup.
Speed and answer quality are different
Timing results do not measure whether an answer is correct. CueFlect uses your saved material to ground its answers; you should review them and speak from your own experience.
Choose where answers happen.
Compare local and cloud processing before choosing a plan.