ArgBench (Few-Shot) - Perspective Assessment: leaderboard

ArgBench (2026) unifies 46 computational-argumentation tasks from 33 existing datasets into five skills. This board is the macro F1 averaged over the perspective assessment tasks under few-shot prompting, times 100; higher is better.

Metric: Macro F1 (0-1, times 100) averaged over the argument perspective assessment (stance, framing and similar) tasks, few-shot prompting; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 11 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-4.166.4#506
2Llama 3.3 70B Instruct62#1394
3Mixtral 8x7B Instruct60.5#4692
4Qwen 3 32B58.8#983
5Qwen 3 4B49.4#2379
6Llama 3.1 8B Instruct46.5#4655
7Mistral-Small-Instruct-240946.3#2026
8Phi-mini-MoE-instruct42.9
9DeepSeek R1 Distill Qwen 32B36.3#1844
10DeepSeek-R1-Distill-Qwen-7B33.4#7976
11Mistral 7B Instruct (v0.3)31.6#6759

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=argbench-few-shot-perspective-assessment · How It Works · Data refreshed daily, snapshot 2026-11-22.