FrontierOR - Quality-Time Efficiency: leaderboard

Quality-time efficiency in percent: the share of large instances over all 180 tasks on which programs that LLMs generate in one shot for large-scale optimization problems from operations research papers, run on hidden large instances against an expert-verified Gurobi reference (FrontierOR, 2026) return a feasible solution within 1 percent of the Gurobi objective and no slower than Gurobi. Higher is better.

Metric: Quality-time efficiency (QTE) times 100 (%): share of large instances over all 180 tasks where the program returns a feasible solution within 1% of the Gurobi objective and no slower than Gurobi, one-shot generation of a complete optimization program from the natural-language problem and instance schemas (a few self-debug iterations allowed, tiny-instance sanity gate), run on the hidden large instances against an expert-verified Gurobi reference; higher is better. Source: arxiv.org. Saturation forecast: Around April 2027. 7 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Claude Opus 4.631#113
2GPT-5.3 Codex26#124
3Gemini 3.1 Pro (Preview)25#99
4DeepSeek R117#482
5Grok 4.2013#301
6Qwen 3 Coder Plus9#720
7Llama 4 Maverick6#1110

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=frontieror-quality-time-efficiency · How It Works · Data refreshed daily, snapshot 2026-11-21.