S2SServiceBench (Decision-Making Handoff): leaderboard
Metric: Level 2 overall score (0-1 printed, shown times 100): unweighted mean over the ten service products (161 expert-selected cases in all) of a schema-constrained handoff report with triggers, timing, constraints and uncertainty-conditioned actions, graded 0-5 on six rubric dimensions by a GPT-5.2 judge and printed on a 0-1 scale; direct prompting with the product artifacts, one response per task; GPT-5.2 is also a subject; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 63.79 | #105 |
| 2 | Claude Opus 4.5 | 50.54 | #79 |
| 3 | Gemini 3 Pro | 37.92 | #77 |
| 4 | Grok 4 | 36.71 | #169 |
| 5 | Qwen 3 VL 32B Instruct | 35.22 | #276 |
| 6 | Llama 4 Maverick Instruct | 20.14 | #439 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=s2sservicebench-decision-making-handoff · How It Works · Data refreshed daily, snapshot 2026-10-11.