Phi-3-medium-128k-instruct [float16] [rev: cae1d42b5577]: benchmark results
Microsoft's Phi-3 Medium instruction model, trained on synthetic and filtered web data and post-trained with supervised fine-tuning and direct preference optimization. This row preserves its context variant and separately evaluated prompt, revision or precision configuration. Provider: Microsoft. Access: Open.
Aggregate Rating 1484 ± 40, rank #2602 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 80.52 | Accuracy (%) (5-shot) | 99.7 |
| Open LLM Leaderboard v1 - MMLU | 76.75 | Accuracy (%) (5-shot) | 98 |
| Open LLM Leaderboard v1 - ARC Challenge | 66.47 | Normalized accuracy (%) (25-shot) | 73.7 |
| Open LLM Leaderboard v1 - HellaSwag | 84.91 | Normalized accuracy (%) (10-shot) | 70.1 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 54.59 | MC2 (%) (0-shot) | 63 |
| Open LLM Leaderboard v1 - WinoGrande | 74.74 | Accuracy (%) (5-shot) | 38.5 |
Interactive version: theaggregate.ai/model?slug=phi-3-medium-128k-instruct-float16-rev-cae1d42b5577 · How It Works · Data refreshed daily, snapshot 2026-11-08.