LiveFact - Classification Accuracy: leaderboard

LiveFact (2026) is a monthly fake-news verification benchmark built from fresh news events: models label generated claims Real, Fake or Ambiguous given evidence available three days before, on, or three days after each event. This board is accuracy in percent against the final verdict when given the evidence from three days after the event (Classification Mode, November 2025 release); higher is better.

Metric: Accuracy (%) in Classification Mode (Real, Fake or Ambiguous against the time-invariant label) with the evidence available three days after the event, LiveFact November 2025; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3 235B A22B 2507 Instruct82.08#625
2GPT-OSS-120B81.81#712
3GPT-5.1 (2025-11-13)81.01#199
4GPT-5.277.32#190
5Qwen 3 30B A3B 2507 Instruct77#1106
6GPT-4o (2024-08-06)73.98#750
7Qwen 3 32B71.86#983
8GPT-4o Mini (2024-07-18)71.22#1501
9Qwen 3 8B69.83#1890
10Llama 3.3 70B Instruct69.76#1394
11GPT-OSS-20B67.42#1188
12Qwen 3 4B 2507 Instruct64.8#2240
13DeepSeek V3.163.73#511
14Kimi K2 (Thinking)54.21
15Llama 3.1 8B Instruct50.2#4655

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

Interactive version: theaggregate.ai/benchmark?slug=livefact-classification-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-22.