Journalistic Bias F1-macro: leaderboard
Journalistic Bias F1-macro evaluates balanced classification quality across media bias categories.
Metric: F1-Macro (%). Source: github.com. Saturation forecast: Around 2033. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 50.24 | #343 |
| 2 | Claude Opus 4.7 | 44.24 | #49 |
| 3 | Gemma 4 31B (IT) | 40.98 | #176 |
| 4 | Llama 3.3 70B Instruct | 35.4 | #534 |
| 5 | GPT-4o Mini | 31.96 | #599 |
| 6 | DeepSeek V3.1 | 30.32 | #264 |
| 7 | GPT-5 (Medium) | 29.13 | #98 (GPT-5) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=journalistic-bias-f1-macro · How It Works · Data refreshed daily, snapshot 2026-10-20.