Ko-WideSearch - Row F1: leaderboard
Metric: F1 (%). Source: arxiv.org. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 53.7 |
| 2 | Claude Opus 4.8 | 52.9 |
| 3 | Claude Opus 4.7 | 51.6 |
| 4 | Claude Opus 4.6 | 48.9 |
| 5 | Gemini 3.1 Pro (Preview) | 45.9 |
| 6 | Gemini 3.1 Flash Lite | 45.9 |
| 7 | DeepSeek V4 Pro | 45 |
| 8 | Gemini 3.5 Flash | 44.1 |
| 9 | Claude Sonnet 4.6 | 43.6 |
| 10 | GPT-5.4 | 41.6 |
| 11 | GLM-5.1 | 34 |
| 12 | GPT-5.4 Mini | 33.3 |
| 13 | Claude Haiku 4.5 | 28.8 |
| 14 | A.X 4.0 | 24.2 |
| 15 | Gemma 4 31B | 23 |
Interactive version: theaggregate.ai/benchmark?slug=ko-widesearch-row-f1 · How It Works · Data refreshed daily, snapshot 2026-09-19.