When2Speak — leaderboard
Metric: 0-shot Macro F1 (self-reported). Source: benchmarklist.com. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o Mini | 56.9 |
| 2 | GPT-4.1 Mini | 49.9 |
| 3 | Qwen 3 8B | 46.6 |
| 4 | Llama 3.1 8B | 33.7 |
| 5 | Llama 3.3 70B Instruct | 32.5 |
Interactive version: theaggregate.ai/benchmark?slug=when2speak · How the rankings work · Data refreshed daily, snapshot 2026-07-22.