| dataset_info: | |
| features: | |
| - name: repo | |
| dtype: string | |
| - name: instance_id | |
| dtype: string | |
| - name: base_commit | |
| dtype: string | |
| - name: patch | |
| dtype: string | |
| - name: test_patch | |
| dtype: string | |
| - name: problem_statement | |
| dtype: string | |
| - name: hints_text | |
| dtype: string | |
| - name: created_at | |
| dtype: string | |
| - name: image_assets | |
| dtype: string | |
| - name: version | |
| dtype: string | |
| - name: FAIL_TO_PASS | |
| dtype: string | |
| - name: PASS_TO_PASS | |
| dtype: string | |
| splits: | |
| - name: dev | |
| num_bytes: 4107331 | |
| num_examples: 102 | |
| - name: test | |
| num_bytes: 893599 | |
| num_examples: 510 | |
| download_size: 1856189 | |
| dataset_size: 5000930 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: dev | |
| path: data/dev-* | |
| - split: test | |
| path: data/test-* | |
| # SWE-bench Multimodal | |
| SWE-bench Multimodal is a dataset of 617 task instances that evalutes Language Models and AI Systems on their ability to resolve real world GitHub issues. | |
| To learn more about the dataset, please visit [our website](https://swebench.com/multimodal). | |
| You can find the leaderboard at SWE-bench's [home page](https://www.swebench.com/#multimodal). |