EuroEval Latvian NLU - Latvian Twitter Sentiment: leaderboard
The "Latvian Twitter Sentiment" subtask of EuroEval Latvian NLU. Latvian-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Matthews correlation coefficient (%). Source: euroeval.com. Saturation forecast: Around February 2027. 233 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 Mini (2026-03-17) (High) | 56.55 | #169 (GPT-5.4 Mini (2026-03-17)) |
| 2 | Gemini 3.1 Flash Lite (Preview) | 54.22 | #194 |
| 3 | Qwen 3 235B A22B 2507 FP8 (Thinking) | 53.84 | #246 (Qwen 3 235B A22B 2507 FP8) |
| 4 | GPT-5 Mini (2025-08-07) | 53.12 | #172 |
| 5 | Qwen 3 30B A3B | 53.11 | #493 |
| 6 | Gemini 3 Flash (Preview) (Thinking) | 53.1 | #82 (Gemini 3 Flash (Preview)) |
| 7 | Mistral Small 3.1 | 52.86 | #610 |
| 8 | Qwen 3 235B A22B 2507 Instruct | 52.23 | #300 |
| 9 | Llama 3.1 70B | 51.96 | #592 |
| 10 | GPT-OSS-20B (Medium) | 51.88 | #518 (GPT-OSS-20B) |
| 11 | Gemini 2.5 Flash (Thinking) | 51.78 | #245 (Gemini 2.5 Flash) |
| 12 | Gemma 3 27B (IT) | 51.39 | #523 |
| 13 | GPT-OSS-20B (High) | 51.26 | #518 (GPT-OSS-20B) |
| 14 | Mistral Small 3.2 | 51.25 | #605 |
| 15 | GPT-5.4 Nano (2026-03-17) (High) | 51.16 | #266 (GPT-5.4 Nano (2026-03-17)) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-latvian-nlu-latvian-twitter-sentiment · How It Works · Data refreshed daily, snapshot 2026-10-20.