EuroEval Swedish NLU - SUC3 — leaderboard

Metric: Named entity recognition Score (%). Source: euroeval.com. 408 models tracked.

Top models

#ModelScore
1GPT-5.4 Mini (High)83.44
2GPT-5.4 Mini (Medium)82.2
3Gemma 4 31B (IT)81.18
4Claude Sonnet 4.5 (Thinking)81
5GPT-580.47
6Gemma 4 26B A4B (IT)79.77
7GPT-5 (High)79.45
8Claude Sonnet 4.5 (Non-reasoning)78.37
9Claude Sonnet 4.677.02
10Gemini 3 Pro (Preview)76.42
11Gemini 2.5 Flash (Thinking)76.3
12Llama 3.1 405B Instruct FP875.9
13GPT-5.275.83
14Grok 4.1 Fast (Reasoning)75.73
15GPT-5 Nano (High)75.73

Interactive version: theaggregate.ai/benchmark?slug=euroeval-swedish-nlu-suc3 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.