EuroEval Icelandic NLU - Hotter and Colder Sentiment — leaderboard

Metric: Sentiment classification Score (%). Source: euroeval.com. 393 models tracked.

Top models

#ModelScore
1Gemini 3 Pro (Preview)65.21
2Gemini 3.1 Flash Lite (Preview)63.57
3Gemma 4 31B (IT)61.64
4GPT-5.4 Mini (Medium)61.64
5GPT-5.260.78
6Grok 2 (1212)60.43
7GPT-4.159.94
8Gemini 2.5 Pro (Preview 03-25)59.15
9Grok 4.1 Fast (Reasoning)59.12
10Grok 358.9
11GPT-558.87
12Gemma 4 26B A4B (IT)58.78
13O3 (2025-04-16)58.74
14Gemini 3 Flash (Preview) (Non-reasoning)58.44
15Gemini 2.5 Flash (Non-reasoning)58.32

Interactive version: theaggregate.ai/benchmark?slug=euroeval-icelandic-nlu-hotter-and-colder-sentiment · How the rankings work · Data refreshed daily, snapshot 2026-07-22.