# Opening Lines Arena leaderboard

external benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelchoice.dev/leaderboards/benchmarks/external-opening-lines-arena

Every model the catalog carries a published Opening Lines Arena value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | external |
| Measure | Relative rating fitted to human pairwise votes |
| Tasks | 50 novel openings: 10 genres with 5 premises each, rated by human votes |
| Difficulty | Fiction writing |
| Published by | [BenchLM](https://benchlm.ai/benchmarks/opening-lines-arena) |

Kaggle's Opening Lines Arena has models write the opening of a novel for 50 premises in 10 genres, and people vote for the better opening.

## Opening Lines Arena ranking

No model in the catalog has a published Opening Lines Arena score. The benchmark is defined by BenchLM, but the catalog carries no value for it yet. An absent value is shown as absent here rather than as a zero.

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
