UnrestSent200K - Few-Shot Accuracy: leaderboard
UnrestSent200K few-shot accuracy under expert prompting. Higher is better. Source: arxiv.org/abs/2609.16997.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o Mini | 76 | #598 |
| 2 | GPT-4.1 Mini | 73 | #363 |
| 3 | GPT-4.1 Nano | 69 | #730 |
| 4 | Qwen 2.5 7B | 61 | #895 |
| 5 | Gemma 3 4B | 59 | #1088 |
| 6 | Qwen 3 4B Instruct | 59 | #661 |
| 7 | Llama 3.1 8B | 58 | #1130 |
| 8 | Qwen 3 4B | 58 | #834 |
| 9 | Qwen 2.5 3B | 55 | #1177 |
| 10 | Gemma 3 1B | 52 | #1505 |
| 11 | Qwen 2.5 1.5B | 49 | #1412 |
| 12 | Llama 3.2 3B | 45 | #1376 |
| 13 | Qwen 2.5 0.5B | 39 | #1536 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=unrestsent200k-few-shot-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-05.