Up to four models
Phi-4 vs Qwen2.5
On the index (GPT-4 = 100), Phi-4 scores 114 and Qwen2.5 112. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
11498–130
WeightsOpen
ReleasedDec 2024
Price in / out$0.070 / $0.14
Context16K
Core tests taken1 of 9
11296–128
WeightsOpen
ReleasedSep 2024
Price in / out$0.36 / $0.40
Context33K
Core tests taken2 of 9
Phi-4 and Qwen2.5 are too close to call (2 points apart, within the uncertainty), ahead on 1 of 1 shared core tests, and costs 4.2× less.
Test by testthird-party results
Every result as a table
| Test | Phi-4 | Qwen2.5 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 56.1% | 49.1% |
| Maths | ||
| AIME-style maths (OTIS mock) | 13.8% | 8.1% |
| MATH (level 5) | 64.9% | 63.2% |
| Agents | ||
| BALROG | 11.6% | 16.2% |
Task length · METR
21 sec
Tests at least two of these models have taken. Bold rows are the core tests.