# SemIf, formerly OpenJev (Qwen3.5-4B, TheoLeeCJ) Benchmark Scores & Performance

> SemIf, formerly OpenJev (Qwen3.5-4B, TheoLeeCJ) by Theodore Lee (TheoLeeCJ) has 2 source-displayable benchmark rows but no public overall score. It remains unranked.

Canonical page: https://benchlm.ai/models/jevbench-semif-qwen3-5-4b

Last updated: September 30, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Theodore Lee (TheoLeeCJ) |
| Source Type | Open Weight |
| Reasoning Type | Non-Reasoning |
| Context Window | Coming soon |
| Overall Score | Coming soon |
| Overall Rank | Unranked |

## Family & Coverage

- Family: SemIf, formerly OpenJev (Qwen3.5-4B, TheoLeeCJ)
- 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) | 47.69 |
| [JevBench 1.5](/benchmarks/jevbench-1-5) | 68.66 |