MoralChoice: leaderboard
MoralChoice: Measures model robustness, truthfulness, calibration, bias, harmfulness, jailbreak resistance, or alignment-relevant behavior.
Metric: MoralChoice Acc (%). Source: trustllmbenchmark.github.io. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4 | 100 | #438 |
| 2 | GLM-4 | 100 | #648 |
| 3 | ChatGPT | 100 | #900 |
| 4 | Llama 3 70B | 99.6 | #631 |
| 5 | Llama 2 70B | 99.1 | #1201 |
| 6 | Mistral 7B | 98.7 | #1397 |
| 7 | Llama 3 8B | 96.9 | #1242 |
| 8 | Llama 2 13B | 96.2 | #1436 |
| 9 | Llama 2 7B | 92 | #1520 |
| 10 | vicuna-13B | 90.5 | #1469 |
| 11 | Vicuna-7B | 59.4 | #1542 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=moralchoice · How It Works · Data refreshed daily, snapshot 2026-10-20.