Phi-3 Mini: benchmark results
Provider: Microsoft. Released 2024-04-23.
Aggregate Rating 1338 ± 43, rank #8001 of 10916 rated models, from 7 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| RubRIX - Dementia Caregiver Queries | 5 | Risk rate (% of the 29 RubRIX audit questions flagged per re | 70 |
| BenSyc - Binary Detection | 52.9 | Binary sycophancy detection macro-F1 (%) (sycophantic versus | 35.7 |
| DiBiMT | 42.58 | Avg Accuracy (%) | 31.6 |
| Wolfram LLM Benchmarking Project | 17.2 | Correct Functionality (%) | 12.1 |
| BenSyc | 19.4 | Five-class conversational alignment classification macro-F1 | 0 |
| Loong | 14.54 | Overall Avg Score (%) | 0 |
| RubRIX - General Caregiver Queries | 12 | Risk rate (% of the 29 RubRIX audit questions flagged per re | 0 |
Interactive version: theaggregate.ai/model?slug=phi-3-mini · How It Works · Data refreshed daily, snapshot 2026-11-08.