# Raw Phi-4 mini direct logits Benchmark Scores & Performance

> Raw Phi-4 mini direct logits by Microsoft / neutral reproduction has 2 source-displayable benchmark rows but no public overall score. It remains unranked.

Canonical page: https://benchlm.ai/models/jevbench-raw-phi-4-mini

Last updated: September 30, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Microsoft / neutral reproduction |
| Source Type | Open Weight |
| Reasoning Type | Non-Reasoning |
| Context Window | Coming soon |
| Overall Score | Coming soon |
| Overall Rank | Unranked |

## Family & Coverage

- Family: Raw Phi-4 mini direct logits
- Variant: decision-system
- Benchmarks covered: 2 of 495
- Coverage note: BenchLM currently has partial benchmark coverage for this model, so the overall score is conservative.

## Reasoning Benchmarks

| Benchmark | Score |
|-----------|-------|
| [JevBench 1.4](/benchmarks/jevbench-1-4) | 37.96 |
| [JevBench 1.5](/benchmarks/jevbench-1-5) | 15.22 |