NovGauge - Pairwise Novelty - Problem Verified F1: leaderboard
Metric: Verified F1 (%). Source: arxiv.org. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 61 |
| 2 | GPT-5.4 Mini | 37.3 |
| 3 | Gemma 4 31B (IT) | 31.7 |
| 4 | Kimi K2.6 | 25.7 |
| 5 | GLM-5.1 | 25.3 |
| 6 | Gemini 3.1 Pro (Preview) | 24.7 |
| 7 | Gemini 3.1 Flash Lite | 22.9 |
| 8 | Qwen 3 32B | 16.2 |
| 9 | Claude Sonnet 4.6 | 13.4 |
| 10 | Qwen 3.6 Plus | 12.3 |
| 11 | DeepSeek V4 Pro | 11.4 |
| 12 | Qwen 3.5 27B | 10.6 |
| 13 | Claude Opus 4.7 | 10.5 |
| 14 | MiniMax-M2.7 | 4.4 |
| 15 | Qwen 3 8B | 3.9 |
Interactive version: theaggregate.ai/benchmark?slug=novgauge-pairwise-novelty-problem-verified-f1 · How It Works · Data refreshed daily, snapshot 2026-09-19.