# Screenplay Entity Benchmark

## Episodes

- `srd part 1 blue revision.pdf`
- `srd part 2 blue revision.pdf`

## Config Summary

| Config | Provider | Mode | Model | Score | Reuse Ratio | Language Violations | Suspicious Pairs | Canon Counts |
| --- | --- | --- | --- | ---: | ---: | ---: | ---: | --- |
| openai_fulltext | openai | fulltext | gpt-5.4 | 85.714 | 0.857 | 0 | 0 | C:9 / L:5 / P:5 |
| gemini_fulltext | gemini | fulltext | gemini-3.1-pro-preview | 82.353 | 0.824 | 0 | 0 | C:11 / L:4 / P:4 |

## Notes

### openai_fulltext
- Score: `85.714`
- Later-episode relink ratio: `0.857`
- Language violations: `0`
- Top suspicious unlinked pairs: none

### gemini_fulltext
- Score: `82.353`
- Later-episode relink ratio: `0.824`
- Language violations: `0`
- Top suspicious unlinked pairs: none