EuroEval Latvian NLU - Latvian Twitter Sentiment — leaderboard

Metric: Sentiment classification Score (%). Source: euroeval.com. 233 models tracked.

Top models

#ModelScore
1GPT-5.4 Mini (High)56.55
2GPT-5.4 Mini (Medium)55.98
3Gemini 3.1 Flash Lite (Preview)54.22
4GPT-5 Mini53.12
5Qwen 3 30B A3B53.11
6Mistral Small 3.152.86
7Qwen 3 235B A22B 2507 Instruct52.23
8GPT-5.4 Nano (Medium)52.2
9Llama 3.1 70B51.96
10GPT-OSS-20B (Medium)51.88
11Gemini 2.5 Flash (Thinking)51.78
12Gemma 3 27B (IT)51.39
13GPT-OSS-20B (High)51.26
14Mistral Small 3.251.25
15GPT-5.4 Nano (High)51.16

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