The Livebench dataset is a time-series dataset related to language modeling. It gathers and processes data from the LiveBench website's GitHub repository and the files used by the live version of the website to ensure the data is up-to-date. The dataset includes information such as question IDs, categories (which are consistently "language"), and release dates for the data. It also contains counts associated with different date ranges and label ranges (e.g., 0.00 - 10.00, 10.00 - 20.00).
Accuracy is the reported evaluation metric for Livebench. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.
Higher is better
Muted rows were not state of the art when published — an earlier or same-year result already scored better.
| Rank | Model | Trust | Score | Year | Links | Fix |
|---|---|---|---|---|---|---|
| 01 | Qwen2.5-Plus | paper | 54.6 | N/A | Paper ↗Code ↗Source ↗ | Looks wrong? |