Text model · Made by Microsoft
Phi-4
Microsoft's Phi-4 small language model — efficient and capable.
- 7.76/10Average score
- #15 of 20Rank among text models
- 30.0sAverage response · #15 fastest
- 9Reviewed results · all time
What the reviews say
- Across 9 reviewed results, Phi-4 averages 7.76/10 and ranks #15 of 20 text models.
- Its relatively weakest area is Relevance (8.31/10, -0.47 against the average model).
- Average response time is 30.0s, #15 fastest of 20 text models.
Scores by dimension
Each dimension is scored 0 to 10 by an independent AI reviewer. The marker shows the average of all ranked text models.
| Dimension | Score | Against the average model |
|---|---|---|
| Accuracy | 7.29 | -0.17 below average |
| Relevance | 8.31 | -0.47 below average |
| Coherence | 8.59 | -0.12 below average |
| Creativity | 5.65 | -0.40 below average |
| Language Quality | 8.65 | -0.08 below average |
Real examples
Recorded comparisons chosen by our team. Excerpts are the opening of the model's actual answer.
Predicting the exact state of AI in two years involves some uncertainty, but we can identify trends and ongoing developments to make educated projections. […]