TERMS-Bench - Surplus Efficiency: leaderboard
Surplus efficiency on TERMS-Bench (2026), a bilateral price-negotiation benchmark in which LLM agents bargain against simulated counterparts with hidden reservation prices, urgency and stance across six behaviour families. The score is the utility the agent earns divided by the width of the zone of possible agreement, averaged over feasible episodes: 1 is the whole surplus, a failed deal counts 0 and a loss-making deal counts below 0. Higher is better.
Metric: Surplus efficiency on feasible episodes (SE+, maximum 1; loss-making agreements are not clipped and count below 0): the agent's utility divided by the zone-of-possible-agreement width, averaged over 1,800 seeded episodes per agent (six counterpart families, three regimes), deterministic decoding; higher is better. Source: arxiv.org. Saturation forecast: Around October 2027. 13 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Claude Opus 4.6 | Max | 0.694 | #113 |
| 2 | GLM-5.1 | 0.686 | #198 | |
| 3 | Claude Opus 4.7 | Max | 0.66 | #85 |
| 4 | Gemma 4 31B (IT) | Thinking | 0.64 | #350 |
| 5 | Gemini 3.1 Pro (Preview) | High | 0.639 | #99 |
| 6 | DeepSeek V4 Pro | Max | 0.618 | #196 |
| 7 | GPT-5.5 | xHigh | 0.606 | #54 |
| 8 | Qwen 3.6 Plus | 0.604 | #238 | |
| 9 | Grok 4.20 | Reasoning | 0.601 | #301 |
| 10 | Kimi K2.6 | 0.597 | #192 | |
| 11 | GPT-5.4 | xHigh | 0.531 | #135 |
| 12 | Seed 2.0 Pro | High | 0.522 | #222 |
| 13 | GPT-4o Mini | 0.189 | #1667 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=terms-bench-surplus-efficiency · How It Works · Data refreshed daily, snapshot 2026-11-21.