HELM ThaiExam - IC — leaderboard
Metric: EM. Source: crfm.stanford.edu. 42 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.5 Sonnet (20240620) | 87.37 |
| 2 | Claude 3 Opus (20240229) | 76.84 |
| 3 | Qwen 2 72B Instruct | 72.63 |
| 4 | Llama 3.1 405B Instruct | 70.53 |
| 5 | Llama 3.1 70B Instruct | 69.47 |
| 6 | Gemini 1.5 Pro (001) | 68.42 |
| 7 | Claude 3 Sonnet (20240229) | 67.37 |
| 8 | Gemini 1.5 Flash (001) | 66.32 |
| 9 | Llama 3 70B Instruct | 65.26 |
| 10 | GPT-4o (2024-05-13) | 63.16 |
| 11 | Qwen1.5 Chat (72B) | 63.16 |
| 12 | GPT-4 Turbo | 61.05 |
| 13 | Claude 3 Haiku (20240307) | 60 |
| 14 | GPT-4o Mini (2024-07-18) | 55.79 |
| 15 | Gemini 1.0 Pro (002) | 46.32 |
Interactive version: theaggregate.ai/benchmark?slug=helm-thaiexam-ic · How the rankings work · Data refreshed daily, snapshot 2026-07-22.