Skip to content

The Aggregate

Unified LLM rankings, daily

Models What's New Trends Benchmarks Skill Maps
How It Works Guesswork Metrics
The Aggregate
Loading data...

HELM SeaHELM - XCOPA (ta) — leaderboard

Metric: EM. Source: crfm.stanford.edu. 21 models tracked.

Top models

#ModelScore
1Claude 3.7 Sonnet (20250219)96.6
2GPT-4o (2024-11-20)96.4
3Gemini 2.0 Flash96
4Llama 3.1 405B Instruct94.6
5Gemini 2.0 Flash Lite94.6
6DeepSeek V393.4
7Llama 3.3 70B Instruct93
8Llama 4 Scout Instruct93
9Llama 3.1 70B Instruct92.6
10GPT-4o Mini (2024-07-18)85.4
11Qwen 2.5 72B Instruct72.2
12Llama 4 Maverick Instruct FP871.6
13Qwen 2 72B Instruct63.4
14Gemma 2 27B (IT)50.4
15Llama 3.2 90B Vision Instruct41.8

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

Built by Mikhail Doroshenko — AI researcher, co-author of Humanity’s Last Exam. Independent project; no affiliation with any AI lab.