ollama
Phi-3 Mini
TL;DR
This page indexes historical reports for Phi-3 Mini. Labels such as “usable” describe the original harness checks, not independently verified correctness. Follow the linked reports for the test conditions.
phi3
3 linked report pages
3 summary lines from local JSON artifacts
- Latest small-eval run averaged 52.4s per question, 4/5 strict passes, and 6.36 tok/s throughput.
- Latest 10Q suite on ollama averaged 111.4s for primary answers with 10/10 usable primary responses.
- Latest 20Q suite on ollama averaged 26.9s for primary answers with 20/20 usable primary responses.