LiveFact - Inference Macro-F1: leaderboard
Metric: Macro-F1 (%) in Inference Mode (label adjusted to what the evidence slice supports) with the evidence available three days after the event, LiveFact November 2025; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.1 | 81.99 |
| 2 | Qwen 3 235B A22B 2507 Instruct | 81.94 |
| 3 | GPT-OSS-120B | 81.31 |
| 4 | GPT-5.2 | 79.57 |
| 5 | Qwen 3 30B A3B 2507 Instruct | 76.4 |
| 6 | GPT-4o (2024-08-06) | 72.98 |
| 7 | Qwen 3 32B | 69.91 |
| 8 | GPT-4o Mini (2024-07-18) | 69.45 |
| 9 | Qwen 3 8B | 68.3 |
| 10 | DeepSeek V3.1 | 66.38 |
| 11 | Qwen 3 4B 2507 Instruct | 64.27 |
| 12 | Llama 3.3 70B Instruct | 63.93 |
| 13 | GPT-OSS-20B | 62.34 |
| 14 | Kimi K2 (Thinking) | 59.15 |
| 15 | Llama 3.1 8B Instruct | 45.14 |
Interactive version: theaggregate.ai/benchmark?slug=livefact-inference-macro-f1 · How It Works · Data refreshed daily, snapshot 2026-10-07.