TemporalBench - MIMIC T1: leaderboard
TemporalBench MIMIC clinical slice for T1 contextual understanding of patient time-series records.
Metric: Accuracy (%). Source: huggingface.co. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | CAMEL (gpt-4o) | 46.81 | |
| 2 | Single LLM (gpt-4o) | 46.81 | |
| 3 | MetaGPT (gpt-4o) | 45.74 | |
| 4 | TimeCopilot (deepseek-chat) | 45.74 | |
| 5 | AgentScope (gpt-4o) | 44.68 | |
| 6 | TimeCopilot (deepseek-v4-pro) | 35.64 | |
| 7 | Single LLM (deepseek-chat) | 33 | |
| 8 | AgentScope (deepseek-chat) | 32.98 | |
| 9 | MetaGPT (deepseek-chat) | 32.98 | |
| 10 | TimeClaw (deepseek-v4-pro) | 32.45 |
Interactive version: theaggregate.ai/benchmark?slug=temporalbench-mimic-t1 · How It Works · Data refreshed daily, snapshot 2026-10-20.