Phi-4-multimodal ranks first on FLEURS.
Microsoft Feb 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 | |
|---|---|---|---|---|
| MMBench-CN v1.1 | 74.9% | #80 of 178 | Official leaderboard | |
| Longform Creative Writing (EQ-Bench) | 22.8 | #117 of 119 | Official leaderboard | |
| POPE | 85.6% | #7 of 7 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HiPhO | 10.6% | #35 of 36 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SnakeBench | 20 | #132 of 188 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| FLEURS | 4.00% | #1 of 6 | Lab model card | |
| SEED-Bench-2-Plus | 68.5% | #26 of 136 | Official leaderboard | |
| MMBench | 83.4% | #30 of 155 | Official leaderboard | |
| CoVoST2 | 39.3 | #3 of 14 | Lab model card | |
| A-Bench | 75.4% | #30 of 116 | Official leaderboard | |
| CCBench (MMBench Chinese Culture) | 55.5% | #49 of 152 | Official leaderboard | |
| MMBench-EN v1.1 | 79.5% | #59 of 178 | Official leaderboard | |
| MMStar | 58.9% | #63 of 173 | Official leaderboard | |
| MM-Vet | 51.9% | #81 of 164 | Official leaderboard | |
| Inter-GPS (Geometry3K) | 48.6% | #1 of 2 | Lab model card | |
| ScienceQA | 97.5% | #1 of 2 | Lab model card | |
| DocVQA | 93.2% | #21 of 36 | Lab model card | |
| Common Voice | 6.80% | #3 of 5 | Lab model card | |
| MMSU | 45% | #3 of 5 | Official leaderboard | |
| CoreCognition | 53.5% | #75 of 119 | Official leaderboard | |
| InfographicVQA | 72.7% | #17 of 26 | Lab model card | |
| VoiceBench | 64.3% | #10 of 15 | Official leaderboard | |
| MMAU v05.15.25 | 65.7% | #4 of 6 | Official leaderboard | |
| TextVQA | 75.6% | #12 of 17 | Lab model card | |
| BLINK | 61.3% | #21 of 29 | Lab model card | |
| ChartQA | 81.4% | #27 of 37 | Lab model card | |
| AVMeme Exam | 45.8% | #3 of 4 | Official leaderboard | |
| OCRBench | 84.4% | #27 of 34 | Lab model card | |
| Audio MC | 15.5% | #15 of 18 | Official leaderboard | |
| MathVista | 62.4% | #59 of 70 | Lab model card | |
| AI2D | 82.3% | #44 of 52 | Lab model card | |
| VoxMem | 20.4% | #8 of 9 | Official leaderboard | |
| MMMU | 55.1% | #92 of 103 | Lab model card | |
| MMMU-Pro | 38.5% | #85 of 90 | Lab model card | |
| Video-MME | 55% | #49 of 51 | Lab model card | |
| MMBench-Video | 0.87 | #28 of 28 | Official leaderboard | |
| MMAU | 55.6% | #15 of 15 | Lab model card | |
| Do You See Me | 24.9% | #5 of 5 | Official leaderboard | |
| AIR-Bench (Audio Instruction Benchmark) | 6.98 | #1 of 1 | Lab model card | |
| Golden3 speech summarization (Microsoft) | 6.28 | #1 of 1 | Lab model card | |
| Open ASR Leaderboard | 6.14% | #1 of 1 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| XSTest | 26.4% | #7 of 16 | Lab model card | |
| RTVLM | 6.39 | #1 of 2 | Lab model card | |
| Harmful content defect rate (Azure AI Evaluation SDK, Microsoft) | 4.00% | #3 of 3 | Lab model card | |
| Microsoft internal multimodal RAI benchmark | 7.96 | #2 of 2 | Lab model card | |
| VLGuard | 8.91 | #2 of 2 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OpenVLM Leaderboard | 64.7 | #58 of 173 | Official leaderboard |
Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.
The benchmarks the lab chose to headline, with the scores it gave.
Every result in the longer documents the lab published with it. 12 of them aren't in the launch post.