Czech Opinion ABSA (10-Shot) - ACOS: leaderboard
Metric: Micro F1 (%) of exact-match aspect-category-opinion-sentiment quad, implicit terms allowed, tuples on the Czech restaurant-review test set (752 sentences, 1,723 tuples), few-shot prompting with the first ten training examples, greedy decoding, open models 4-bit quantized; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.3 70B Instruct [4bit] | 33.76 | #473 |
| 2 | GPT-4o Mini | 33.05 | #588 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=czech-opinion-absa-10-shot-acos · How It Works · Data refreshed daily, snapshot 2026-10-11.