phi-2-upscaled-4B-instruct-v0.1: benchmark results
Provider: Microsoft. Access: Open.
Aggregate Rating 1109 ± 42, rank #10618 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 40.92 | MC2 (%) (0-shot) | 15.6 |
| Open LLM Leaderboard v1 - GSM8K | 0.76 | Accuracy (%) (5-shot) | 14 |
| Open LLM Leaderboard v1 - MMLU | 26.8 | Accuracy (%) (5-shot) | 13.8 |
| Open LLM Leaderboard v1 - HellaSwag | 28.68 | Normalized accuracy (%) (10-shot) | 4.8 |
| Open LLM Leaderboard v1 - WinoGrande | 50.59 | Accuracy (%) (5-shot) | 3.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 22.95 | Normalized accuracy (%) (25-shot) | 2.2 |
Interactive version: theaggregate.ai/model?slug=phi-2-upscaled-4b-instruct-v0-1 · How It Works · Data refreshed daily, snapshot 2026-11-08.