PatRe - Office Action Decision (OA-RS): leaderboard
Metric: Decision accuracy (%): the generated office action's decision (allowance, non-final or final rejection, Ex parte Quayle) matches the examiner's, office action generation under retrieval simulated: the model receives a noisy pool of BM25-retrieved references mixed with ground-truth ones, over PatRe (480 recent USPTO patent examination histories across all eight IPC sections, with office actions, rebuttals, claim versions and cited references), temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around February 2028. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Flash | 52.8 |
| 2 | GPT-5 Mini | 52.7 |
| 3 | Qwen 3.5 27B | 50.5 |
| 4 | Qwen 3.5 9B | 43.4 |
| 5 | DeepSeek V3.2 | 42.2 |
| 6 | Gemma 3 12B (IT) | 41.6 |
| 7 | Llama 3.1 8B Instruct | 39.9 |
| 8 | Gemma 3 27B (IT) | 38.1 |
| 9 | GPT-4o Mini | 36.3 |
| 10 | Llama 3.3 70B Instruct | 21.8 |
Interactive version: theaggregate.ai/benchmark?slug=patre-office-action-decision-oa-rs · How It Works · Data refreshed daily, snapshot 2026-10-07.