CalBench - Varied Costs - Coordination: leaderboard
Fraction of requested meetings completed. Canonical mixed-agent calendar negotiation, paper v3 Table 1. Uniform and varied displacement costs are separate settings; 45 or 135 evaluated agent-game seats per model, not homogeneous teams. Means retain original units. This series reports Coordination (%). Source: https://arxiv.org/html/2605.09823v3.
Metric: Coordination (%). Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4 Pro | 84.44 | #112 |
| 2 | Claude Sonnet 4.6 | 84.2 | #94 |
| 3 | Gemini 3.1 Pro (Preview) | 81.73 | #64 |
| 4 | Qwen 3.6 Plus | 79.26 | #124 |
| 5 | Gemini 3 Flash (Preview) | 78.52 | #90 |
| 6 | Llama 4 Maverick Instruct | 77.04 | #467 |
| 7 | GPT-5.4 Mini | 76.79 | #228 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=calbench-varied-costs-coordination · How It Works · Data refreshed daily, snapshot 2026-11-05.