NovGauge - Pairwise Novelty - Problem F1: leaderboard
Metric: F1 (%). Source: arxiv.org. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Flash Lite | 89.2 |
| 2 | Gemini 3.1 Pro (Preview) | 77.1 |
| 3 | Qwen 3.5 27B | 76.7 |
| 4 | Kimi K2.6 | 76.5 |
| 5 | Qwen 3.6 Plus | 74.9 |
| 6 | MiniMax-M2.7 | 73.3 |
| 7 | GPT-5.5 | 69.4 |
| 8 | GLM-5.1 | 69.3 |
| 9 | Gemma 4 31B (IT) | 66.2 |
| 10 | Claude Opus 4.7 | 66.2 |
| 11 | Claude Sonnet 4.6 | 66.1 |
| 12 | DeepSeek V4 Pro | 60.3 |
| 13 | DeepSeek V4 Flash | 60 |
| 14 | GPT-5.4 Mini | 56.7 |
| 15 | Qwen 3 32B | 56.6 |
Interactive version: theaggregate.ai/benchmark?slug=novgauge-pairwise-novelty-problem-f1 · How It Works · Data refreshed daily, snapshot 2026-09-19.