JointDW - Templates: leaderboard
The JointDW joint-support question on nine textbook lot-sizing, network-design and facility-location programs, three samples each; the score is hit@1. The paper finds the answers here follow the constraint names: with names permuted or anonymised the best model finds no joint support.
Metric: hit@1: share of answers whose first predicted joint support is true, nine lot-sizing, network-design and facility-location templates, three samples each (0-1). Source: arxiv.org. Saturation forecast: Around February 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6.1 Sol (Medium) | 1 | #10 (GPT-6.1 Sol) |
| 2 | Claude Sonnet 5.5 | 0.7 | #19 |
| 3 | Grok 4.20 (Reasoning) | 0.67 | #297 (Grok 4.20) |
| 4 | Grok 4.20 (Non-reasoning) | 0.33 | #297 (Grok 4.20) |
| 5 | DeepSeek V4 Pro (Non-reasoning) | 0.3 | #188 (DeepSeek V4 Pro) |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=jointdw-templates · How It Works · Data refreshed daily, snapshot 2026-11-13.