CIBuzzBench - Equivalent Selection (English Prompt): leaderboard
Metric: Macro F1 (%; four-way choice of the English equivalent against literal, cultural-mismatch and pragmatic-neighbor distractors, mean over five option-shuffle seeds; English instructions; 3,001 Chinese internet buzzwords; zero-shot, temperature 0; the Chinese buzzword is the input under both prompt languages). Source: arxiv.org. Saturation forecast: Around December 2026. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 88.03 |
| 2 | GPT-5.5 | 87.12 |
| 3 | Qwen 3.7 Max | 82.38 |
| 4 | GLM-5.2 | 82.27 |
| 5 | Claude Opus 4.8 | 79.94 |
| 6 | DeepSeek V4 Pro | 79.66 |
| 7 | Qwen 3 14B | 56.68 |
| 8 | Qwen 3 8B | 54.59 |
| 9 | Qwen 3 4B | 45.18 |
Interactive version: theaggregate.ai/benchmark?slug=cibuzzbench-equivalent-selection-english-prompt · How It Works · Data refreshed daily, snapshot 2026-09-26.