HELM SeaHELM - XNLI (th): leaderboard
Metric: EM. Source: crfm.stanford.edu. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.1 405B Instruct | 70 |
| 2 | GPT-4o (2024-11-20) | 69.4 |
| 3 | Claude 3.7 Sonnet (20250219) | 66.3 |
| 4 | Qwen 2 72B Instruct | 63.5 |
| 5 | Gemini 2.0 Flash | 63 |
| 6 | Llama 3.3 70B Instruct | 60.4 |
| 7 | Llama 3.1 70B Instruct | 58.9 |
| 8 | Gemini 2.0 Flash Lite | 57.2 |
| 9 | Qwen 2.5 72B Instruct | 56.9 |
| 10 | Llama 3.2 90B Vision Instruct | 56.1 |
| 11 | GPT-4o Mini (2024-07-18) | 56 |
| 12 | Claude 3.5 Haiku (20241022) | 55.1 |
| 13 | Llama 4 Maverick Instruct FP8 | 50.7 |
| 14 | DeepSeek V3 | 50.5 |
| 15 | DeepSeek R1 Distill Llama 8B | 33.3 |
Interactive version: theaggregate.ai/benchmark?slug=helm-seahelm-xnli-th · How It Works · Data refreshed daily, snapshot 2026-09-05.