PERSPECTRA - Polarity Check: leaderboard
Metric: Accuracy (%) of the binary dominant stance (pro or con, the majority of the included expansions' stances) of a paragraph concatenating several expanded arguments of one topic (PERSPECTRA: Kialo debate opinions (100 topics, 762 pro and con opinions) each expanded into five Reddit-style arguments by GPT-4o; 500 sampled evaluation items per task, gold labels derived from the Kialo opinion identifiers and stances); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 76.4 | #333 |
| 2 | GPT-4o Mini | 72.8 | #588 |
| 3 | Qwen 3 32B | 67.6 | #424 |
| 4 | DeepSeek R1 Distill Qwen 32B | 67.2 | #640 |
| 5 | QwQ-32B | 66.4 | #410 |
| 6 | Qwen 3 8B | 64 | #667 |
| 7 | Qwen 2.5 7B Instruct | 58.6 | #846 |
| 8 | DeepSeek-R1-Distill-Qwen-7B | 54.8 | #1400 |
| 9 | DeepSeek R1 Distill Llama 8B | 54.8 | #1282 |
| 10 | Falcon3-7B-Instruct | 51 | #1284 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=perspectra-polarity-check · How It Works · Data refreshed daily, snapshot 2026-10-11.