Large Language Models Lack Temporal Awareness: leaderboard
The existing methods for evaluating the medical knowledge of Large Language Models (LLMs) are largely based on atemporal examination-style benchmarks, while in reality, medical knowledge is inherently dynamic and continuously evolves as new evidence emerges and treatments are approved. Consequently, evaluating medical knowledge without a temporal context may provide an incomplete assessment of whe
Metric: Accuracy (self-reported). Source: benchmarklist.com. Saturation forecast: Around May 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 | 71.11 | #248 |
| 2 | GPT-5 | 70.69 | #98 |
| 3 | MedGemma-4B | 60.1 | #746 |
| 4 | Qwen 2.5 7B | 59.64 | #922 |
| 5 | Gemma 3 4B | 58.16 | #1089 |
| 6 | MedGemma 1.5 4B | 55.43 | #789 |
| 7 | Llama 3.1 8B | 54.92 | #1151 |
| 8 | GPT-OSS-20B | 54.05 | #518 |
| 9 | Llama 3.2 3B | 53.63 | #1396 |
| 10 | Qwen 2.5 14B | 46.32 | #725 |
| 11 | Qwen 3 4B | 46.28 | #829 |
| 12 | GPT-4o | 44.95 | #343 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=large-language-models-lack-temporal-awareness · How It Works · Data refreshed daily, snapshot 2026-10-20.