Phi-4-Reasoning-plus.
Microsoft Apr 2025 Open weights Small
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Arena-Hard-Auto | 79% | #19 of 39 | Model card | |
| FlenQA | 97.9% | #1 of 2 | Model card | |
| IFEval | 84.9% | #75 of 126 | Model card | |
| MMLU-Pro | 76% | #98 of 151 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| AIME 2024 | 81.3% | #37 of 87 | Model card | |
| Misguided Attention | 48.4% | #21 of 45 | Official leaderboard | |
| Omni-MATH | 81.9% | #2 of 4 | Model card | |
| AIME 1983-2024 | 89.4% | #1 of 2 | Model card | |
| TSP-Opt | 42.6% | #1 of 2 | Model card | |
| GPQA Diamond | 69.3% | #176 of 284 | Model card | |
| AIME 2025 | 78% | #95 of 143 | Model card | |
| HMMT 2025 | 53.6% | #47 of 49 | Model card | |
| 3SAT-Search | 70.9% | #2 of 2 | Model card | |
| BA-Calendar | 65.6% | #2 of 2 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HumanEval+ | 92.3% | #5 of 27 | Model card | |
| LiveCodeBench | 53.1% | #40 of 73 | Model card | |
| Codeforces rating | 1,723 | #32 of 41 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SpatialEval Maze-Nav | 53.4% | #2 of 2 | Model card | |
| SpatialEval Spatial-Map | 73.3% | #2 of 2 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ToxiGen (discriminative) | 85% | #2 of 2 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| PhiBench | 74.2% | #1 of 3 | Model card |
The benchmarks the lab chose to headline, with the scores it gave.
Every result in the longer documents the lab published with it. 9 of them aren't in the launch post.