ArgBench (Chain-of-Thought) - Perspective Assessment: leaderboard
Metric: Macro F1 (0-1, times 100) averaged over the argument perspective assessment (stance, framing and similar) tasks, chain-of-thought prompting; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4.1 | 60.3 |
| 2 | Qwen 3 32B | 55.2 |
| 3 | Llama 3.3 70B Instruct | 53.1 |
| 4 | Mixtral 8x7B Instruct (v0.1) | 52.3 |
| 5 | Llama 3.1 8B Instruct | 49.2 |
| 6 | Qwen 3 4B | 48.4 |
| 7 | DeepSeek R1 Distill Qwen 32B | 47.8 |
| 8 | Mistral-Small-Instruct-2409 | 47.4 |
| 9 | DeepSeek-R1-Distill-Qwen-7B | 42.2 |
| 10 | Mistral 7B Instruct (v0.3) | 38.8 |
Interactive version: theaggregate.ai/benchmark?slug=argbench-chain-of-thought-perspective-assessment · How It Works · Data refreshed daily, snapshot 2026-10-07.