ORAgentBench: leaderboard
Large language models are increasingly deployed as autonomous agents for multi-step tasks in executable environments, yet their ability to perform realistic operations research (OR) work remains unclear. Existing OR evaluations often decouple modeling from solving, rely on pre-formalized or text-only instances, and rarely test the full workflow from operational artifacts to validated decisions. In
Metric: Pass Rate All (self-reported). Source: benchmarklist.com. Saturation forecast: Around August 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 35.51 | #79 |
| 2 | Claude Opus 4.6 | 34.58 | #63 |
| 3 | GPT-5.3 Codex | 32.71 | #69 |
| 4 | DeepSeek V4 Pro | 27.1 | #102 |
| 5 | Kimi K2.6 | 27.1 | #104 |
| 6 | DeepSeek V4 Flash | 26.17 | #137 |
| 7 | GPT-5.4 Mini | 26.17 | #212 |
| 8 | GLM-5.1 | 26.17 | #97 |
| 9 | Claude Sonnet 4.6 | 25.23 | #90 |
| 10 | Qwen 3.6 Plus | 22.43 | #120 |
| 11 | GLM-5 | 22.43 | #141 |
| 12 | MiMo-V2.5-Pro | 22.43 | #158 |
| 13 | Qwen 3.5 Plus | 20.56 | #125 |
| 14 | MiniMax-M2.7 | 14.02 | #258 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=oragentbench · How It Works · Data refreshed daily, snapshot 2026-10-20.