SpeakerSleuth - Discrimination: leaderboard
Metric: Classification accuracy (%; 606 three-candidate items). Source: arxiv.org. Saturation forecast: Estimated already saturated. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 81.5 |
| 2 | Gemini 2.5 Flash | 75.6 |
| 3 | Gemini 2.5 Flash Lite | 46.5 |
| 4 | Phi-4 Multimodal Instruct | 39.9 |
| 5 | Gemma 3n E4B | 37.1 |
Interactive version: theaggregate.ai/benchmark?slug=speakersleuth-discrimination · How It Works · Data refreshed daily, snapshot 2026-09-25.