MultiCW (CoT): leaderboard
Metric: Accuracy (%) of zero-shot binary check-worthiness classification (claim worth fact-checking or not) on the 18,444-sample MultiCW test split, balanced across 16 languages, two writing styles (noisy social media and structured news) and the two classes, with a chain-of-thought prompt built from the CLEF annotation guidelines (CoT_CLEF_on_Q), decisions extracted from the response with a strict yes/no fallback prompt; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Haiku | 75 | #553 |
| 2 | Claude 3.7 Sonnet | 72 | #241 |
| 3 | Llama 3.3 70B Instruct | 71 | #520 |
| 4 | GPT-4o | 69 | #333 |
| 5 | GPT-4o Mini | 69 | #588 |
| 6 | Mistral Nemo 12B | 66 | #1239 |
| 7 | GPT-4.1 Mini | 65 | #346 |
| 8 | GPT-4.1 | 65 | #240 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=multicw-cot · How It Works · Data refreshed daily, snapshot 2026-10-11.