HELM SeaHELM - Wisesight — leaderboard
Metric: Macro F1 score. Source: crfm.stanford.edu. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 55.49 |
| 2 | Claude 3.7 Sonnet (20250219) | 55 |
| 3 | GPT-4o (2024-11-20) | 54.76 |
| 4 | Gemini 2.0 Flash Lite | 54.73 |
| 5 | Gemini 2.0 Flash | 54.52 |
| 6 | Qwen 2 72B Instruct | 53.54 |
| 7 | GPT-4o Mini (2024-07-18) | 52.29 |
| 8 | DeepSeek V3 | 50.7 |
| 9 | Llama 4 Maverick Instruct FP8 | 49.7 |
| 10 | Claude 3.5 Haiku (20241022) | 48.25 |
| 11 | Llama 4 Scout Instruct | 46.58 |
| 12 | Llama 3.1 405B Instruct | 43.96 |
| 13 | Llama 3.1 70B Instruct | 39.63 |
| 14 | Llama 3.3 70B Instruct | 38.76 |
| 15 | Llama 3.2 90B Vision Instruct | 33.43 |
Interactive version: theaggregate.ai/benchmark?slug=helm-seahelm-wisesight · How the rankings work · Data refreshed daily, snapshot 2026-07-22.