TemporalBench - CausalChambers T4: leaderboard
TemporalBench Causal Chambers slice for T4 event-conditioned forecasting under causal changes.
Metric: Accuracy (%). Source: huggingface.co. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | TimeClaw (deepseek-v4-pro) | 51.33 | |
| 2 | CAMEL (deepseek-chat) | 48 | |
| 3 | AgentScope (deepseek-chat) | 42 | |
| 4 | MetaGPT (deepseek-chat) | 40 | |
| 5 | AgentScope (gpt-4o) | 32 | |
| 6 | Single LLM (deepseek-chat) | 29.33 | |
| 7 | CAMEL (gpt-4o) | 26.67 | |
| 8 | Single LLM (gpt-4o) | 26 | |
| 9 | MetaGPT (gpt-4o) | 16 | |
| 10 | TimeSeries Scientist (deepseek-chat) | 2.67 |
Interactive version: theaggregate.ai/benchmark?slug=temporalbench-causalchambers-t4 · How It Works · Data refreshed daily, snapshot 2026-10-20.