VIVID - Topic Classification: leaderboard
Exact-match classification of each Vietnamese idiom into one of seven semantic themes in VIVID (2026), three-shot. Accuracy in percent, higher is better.
Metric: Accuracy (%; exact-match choice of one of seven semantic themes for each idiom or proverb, three-shot, LM Evaluation Harness). Source: arxiv.org. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 53.5 | #452 |
| 2 | GPT-4o | 52.4 | #741 |
| 3 | Llama 4 Scout Base | 45.9 | #1890 |
| 4 | Qwen 3 14B | 44.7 | #1235 |
| 5 | Llama-SEA-LION-v3-8B-IT | 31.9 | #2317 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=vivid-topic-classification · How It Works · Data refreshed daily, snapshot 2026-11-12.