NovGauge - Multi-Paper Grouping - Method F1: leaderboard
Metric: F1 (%). Source: arxiv.org. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Flash Lite | 70.7 |
| 2 | GPT-5.5 | 64.6 |
| 3 | Kimi K2.6 | 63.4 |
| 4 | Claude Opus 4.7 | 63.3 |
| 5 | Claude Sonnet 4.6 | 61.3 |
| 6 | Gemini 3.1 Pro (Preview) | 58.7 |
| 7 | DeepSeek V4 Pro | 55.3 |
| 8 | Qwen 3.6 Plus | 53.4 |
| 9 | GLM-5.1 | 53.2 |
| 10 | GPT-5.4 Mini | 50.4 |
| 11 | MiniMax-M2.7 | 49.4 |
| 12 | DeepSeek V4 Flash | 48.1 |
Interactive version: theaggregate.ai/benchmark?slug=novgauge-multi-paper-grouping-method-f1 · How It Works · Data refreshed daily, snapshot 2026-09-19.