PIJ - Sentence Prediction: leaderboard
F1 for sentence prediction on the PIJ human-comparison subset of 50 samples per task, including Unknown labels. These are the human-comparison results rather than the full 2,500-case headline evaluation.
Metric: F1 (%, 50-sample human evaluation, include Unknown). Source: arxiv.org. Saturation forecast: Around January 2028. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 65.87 | #103 |
| 2 | GPT-5.4 | 57.73 | #83 |
| 3 | Claude Sonnet 4.6 | 51.94 | #94 |
| 4 | Qwen 3 14B | 44.47 | #542 |
| 5 | DeepSeek V4 Pro (Non-reasoning) | 44.39 | #112 (DeepSeek V4 Pro) |
| 6 | GPT-5.4 Mini | 43.48 | #228 |
| 7 | Qwen 3 Max | 39.7 | #213 |
| 8 | DeepSeek V4 Pro (Reasoning) | 37.93 | #112 (DeepSeek V4 Pro) |
| 9 | Llama 3.3 70B | 35.25 | #578 |
| 10 | Qwen 3 8B | 31.32 | #678 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=pij-sentence-prediction · How It Works · Data refreshed daily, snapshot 2026-11-05.