Microsoft — AI model benchmarks
AI models from Microsoft. 26 models tracked, 26 with a unified ELO rating. Current best: Phi-Line_14B (ELO 1550, rank #620 overall). Most recent tracked release: 2025-04-30.
Rated models
| Rank | Model | Provider | Unified ELO | Benchmarks |
|---|---|---|---|---|
| 620 | Phi-Line_14B | Microsoft | 1550 ± 31 | 8 |
| 725 | Phi-3.5-MoE-instruct | Microsoft | 1520 ± 15 | 40 |
| 764 | WizardLM-2 8x22B | Microsoft | 1511 ± 18 | 37 |
| 859 | Phi-lthy4 | Microsoft | 1485 ± 36 | 8 |
| 867 | Phi-4 | Microsoft | 1482 ± 6 | 387 |
| 946 | Phi-3-medium-4k-instruct-abliterated-v3 | Microsoft | 1463 ± 54 | 7 |
| 993 | Phi-3-medium-128k-instruct | Microsoft | 1451 ± 16 | 44 |
| 999 | Phi-3-small-8k-instruct | Microsoft | 1450 ± 16 | 42 |
| 1000 | Phi-3-small-128k-instruct | Microsoft | 1450 ± 16 | 28 |
| 1028 | Phi-3-medium-4k-instruct | Microsoft | 1444 ± 13 | 77 |
| 1061 | Phi-3.5-mini-instruct | Microsoft | 1436 ± 12 | 120 |
| 1107 | phi-4-14B | Microsoft | 1425 ± 93 | 16 |
| 1147 | Phi-4 Mini Instruct | Microsoft | 1416 ± 9 | 276 |
| 1190 | Phi-3-mini-4K-instruct-cpo-simpo | Microsoft | 1407 ± 10 | 43 |
| 1259 | Phi-3 Mini 4K Instruct | Microsoft | 1393 ± 15 | 76 |
Interactive version: theaggregate.ai/company?slug=microsoft · How the rankings work · Data refreshed daily, snapshot 2026-07-22.