PickMeBench: leaderboard
1v1 negotiation benchmark where frontier models argue over which subscription to cancel. Win rate across all pairwise matchups measures persuasion and strategic reasoning.
Metric: Win Rate (%). Source: pick-me-bench-site.vercel.app. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 78 | #63 |
| 2 | Grok 4.20 0309 (Reasoning) | 69.3 | #112 (Grok 4.20 0309) |
| 3 | GPT-5.4 | 48.7 | #79 |
| 4 | Kimi K2.5 | 33.3 | #144 |
| 5 | Gemini 3.1 Pro (Preview) | 32 | #58 |
| 6 | DeepSeek Reasoner | 23.3 | #242 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=pickmebench · How It Works · Data refreshed daily, snapshot 2026-10-20.