TemporalBench - CausalChambers T1: leaderboard
TemporalBench Causal Chambers slice for T1 contextual understanding of controlled causal time-series environments.
Metric: Accuracy (%). Source: huggingface.co. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | TimeSeries Scientist (deepseek-chat) | 28.67 | |
| 2 | TimeSeries Scientist (gpt-4o) | 28.67 | |
| 3 | TimeClaw (deepseek-v4-pro) | 22 | |
| 4 | TimeCopilot (deepseek-chat) | 22 | |
| 5 | TimeCopilot (deepseek-v4-pro) | 17.33 | |
| 6 | AgentScope (deepseek-chat) | 16 | |
| 7 | MetaGPT (deepseek-chat) | 15.33 | |
| 8 | Single LLM (deepseek-chat) | 15.33 | |
| 9 | Single LLM (gpt-4o) | 13.33 | |
| 10 | CAMEL (deepseek-chat) | 12.67 |
Interactive version: theaggregate.ai/benchmark?slug=temporalbench-causalchambers-t1 · How It Works · Data refreshed daily, snapshot 2026-10-20.