HELM TORR - Numeric Nlg - Robustness: leaderboard
Metric: Robustness (0-100). Source: crfm.stanford.edu. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.5 Sonnet (20241022) | 92.96 |
| 2 | GPT-4o Mini (2024-07-18) | 92.73 |
| 3 | GPT-4o (2024-11-20) | 92.16 |
| 4 | Claude 3.5 Haiku (20241022) | 91.68 |
| 5 | Llama 3.1 8B Instruct | 91.15 |
| 6 | Gemini 1.5 Flash (002) | 90.51 |
| 7 | Gemini 1.5 Pro (002) | 88.21 |
| 8 | Mistral 7B Instruct (v0.3) | 87.12 |
| 9 | Llama 3.1 70B Instruct | 86.88 |
| 10 | Qwen 2 72B Instruct | 86.74 |
| 11 | DeepSeek V3 | 85.87 |
| 12 | Llama 3.1 405B Instruct | 85.08 |
Interactive version: theaggregate.ai/benchmark?slug=helm-torr-numeric-nlg-robustness · How It Works · Data refreshed daily, snapshot 2026-09-19.