HELM SeaHELM - IndicSentiment — leaderboard

Metric: Macro F1 score. Source: crfm.stanford.edu. 21 models tracked.

Top models

#ModelScore
1Gemini 2.0 Flash98.59
2Llama 3.1 405B Instruct98.3
3Gemini 2.0 Flash Lite98.24
4Llama 3.3 70B Instruct98.2
5Llama 4 Scout Instruct98.05
6Llama 3.2 90B Vision Instruct98
7GPT-4o Mini (2024-07-18)97.8
8Claude 3.7 Sonnet (20250219)97.75
9Llama 3.1 70B Instruct97.7
10DeepSeek V397.5
11GPT-4o (2024-11-20)97.3
12Qwen 2.5 72B Instruct93.89
13Claude 3.5 Haiku (20241022)93.48
14Qwen 2 72B Instruct75.29
15Gemma 2 27B (IT)45.34

Interactive version: theaggregate.ai/benchmark?slug=helm-seahelm-indicsentiment · How the rankings work · Data refreshed daily, snapshot 2026-07-22.