Phi-1.5: benchmark results

Microsoft's Phi-1.5, a 1.3B transformer base model trained on the same data sources as Phi-1 plus new synthetic NLP texts, excluding generic web-crawl data and without RLHF fine-tuning. Provider: Microsoft. Released 2023-09-11. Access: Open.

Aggregate Rating 1217 ± 15, rank #9200 of 10318 rated models, from 189 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LA Leaderboard - WNLI Basque61.97Accuracy (%)97.8
MMLU-by-task - College Mathematics41Accuracy (%)97.2
LA Leaderboard - WiC Basque50.29Accuracy (%)89
MMLU-by-task - Abstract Algebra34Accuracy (%)82
AI Energy Score (Text Generation)5Energy Score (1-5)79.5
MMLU-by-task - College Computer Science46Accuracy (%)75.6
MMLU-by-task - Machine Learning38.39Accuracy (%)75.4
LA Leaderboard - HumorQA58.82Accuracy (%)70.6
MMLU-by-task - Electrical Engineering48.28Accuracy (%)68.6
MMLU-by-task - College Physics27.45Accuracy (%)62.9
MMLU-by-task - Business Ethics52Accuracy (%)61.9
MMLU-by-task - Moral Disputes54.91Accuracy (%)61.6

Interactive version: theaggregate.ai/model?slug=phi-1-5 · How It Works · Data refreshed daily, snapshot 2026-11-14.