| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Legal Reasoning | PRBench Legal-Hard | GPT-5 Grader Score0.377 | 4 | |
| Legal Reasoning | PRBench Legal | GPT-5 Grader Score0.482 | 4 | |
| Legal reasoning | PRBench-Legal 500 prompts | GPT-5 Grader Score44.1 | 4 | |
| Generative Performance | PRBench | Pearson r0.8 | 1 |