Phi-4
| Released | 2024-12-12 |
|---|---|
| Price per 1M | Not listed |
| Overall score | 35.0 · rank 217 |
| Provider | Microsoft |
| Weights | Open weights |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 35.0 | 217 | 7 |
| Overall: Coding | 35.5 | 206 | 4 |
| Overall: Reasoning | 39.2 | 198 | 4 |
| Overall: Agentic work | 34.5 | 221 | 4 |
| Agents | 33.1 | 178 | 2 |
| Reasoning | 40.2 | 194 | 2 |
| Math | 38.0 | 171 | 2 |
| General | 28.1 | 184 | 1 |
Evidence
Agents 33.1 ±3.2
Reasoning 40.2 ±3.5
| Benchmark | Result | Source | Date |
|---|---|---|---|
| GPQA Diamond | 56.1% | Epoch AI | 2025-01-31 |
| CritPt | 0.0% | Epoch AI | 2026-11-11 |
Math 38.0 ±2.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MATH Level 5 | 64.9% | Epoch AI | 2025-01-31 |
| OTIS Mock AIME 2024-2025 | 13.8% | Epoch AI | 2025-02-25 |
General 28.1 ±4.1
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LMArena Text | 1,256 | LMArena | 2026-11-09 |
Names matched to this model
phi-4, phi-4-reasoning, phi4-14b-q8