TemporalBench - MIMIC T3: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | TimeClaw (deepseek-v4-pro) | 56.86 |
| 2 | TimeCopilot (deepseek-v4-pro) | 52 |
| 3 | Single LLM (deepseek-chat) | 39.14 |
| 4 | Single LLM (gpt-4o) | 36.61 |
| 5 | TimeCopilot (deepseek-chat) | 35.42 |
| 6 | CAMEL (gpt-4o) | 30.14 |
| 7 | MetaGPT (gpt-4o) | 28.97 |
| 8 | TimeSeries Scientist (gpt-4o) | 28.87 |
| 9 | TimeSeries Scientist (deepseek-chat) | 25.94 |
| 10 | MetaGPT (deepseek-chat) | 25.79 |
Interactive version: theaggregate.ai/benchmark?slug=temporalbench-mimic-t3 · How It Works · Data refreshed daily, snapshot 2026-09-05.