SCHEDBench (Full-Variation): leaderboard

Metric: Feasibility rate (%; share of SCHEDBench natural-language scheduling instances whose model output satisfies all hard constraints, verified by OR-Tools CP-SAT; six canonical job-shop, project-scheduling, nurse-rostering and timetabling families; greedy decoding, zero-shot, no external solver; all three surface-form axes (themed names, paraphrased templates, constraint reordering) active). Source: arxiv.org. Saturation forecast: Around September 2028. 13 models tracked.

Top models

#ModelScore
1Gemini 3 Flash (Preview)29.5
2Qwen 3.5 27B25
3Qwen 3.5 397B A17B24.3
4Qwen 3.5 122B A10B14.3
5Claude Sonnet 4.612.8
6Claude Opus 4.68.9
7Claude Haiku 4.54.9
8Llama 3.3 70B4.9
9Llama 4 Maverick4.5
10Gemini 3.1 Flash Lite3.9
11GPT-5.4 (2026-03-05)3.7

Interactive version: theaggregate.ai/benchmark?slug=schedbench-full-variation · How It Works · Data refreshed daily, snapshot 2026-09-26.