CESBench - Scenario: leaderboard
The 63 scenario items of CESBench (Zhou et al., 2026): open-ended diagnosis of a concrete cryptographic-engineering situation with platform, measurement and trace details. Scored in percent as the mean of three to five expert rubric dimensions judged by Qwen3.5-397B; higher is better.
Metric: Scenario score (%; the 63 open-ended engineering diagnosis items, mean of 3 to 5 expert rubric dimensions judged 0, 0.5 or 1; 380 expert-written items on cryptographic engineering security for IoT devices; zero-shot, temperature 0, no tools, provider default reasoning mode; judgment and scenario answers scored by a Qwen3.5-397B judge from outside the evaluated set). Source: arxiv.org. Saturation forecast: Around February 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Luna | 88.4 | #117 |
| 2 | GLM-5.2 | 84.2 | #137 |
| 3 | Kimi K2.6 | 82.2 | #173 |
| 4 | Gemini 3.7 Flash | 81.9 | #32 |
| 5 | DeepSeek V4 Pro | 81.4 | #181 |
| 6 | MiniMax-M2.5 | 71.9 | #590 |
| 7 | DeepSeek V3 | 67.5 | #705 |
| 8 | GLM-4 32B (0414) | 59.1 | #1407 |
| 9 | Llama 4 Maverick | 57.1 | #1105 |
| 10 | Ling-flash-2.0 | 55.9 | #1352 |
| 11 | Hunyuan A13B-Instruct | 47.2 | #2021 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=cesbench-scenario · How It Works · Data refreshed daily, snapshot 2026-11-12.