TemporalBench - CausalChambers T3: leaderboard
TemporalBench Causal Chambers slice for T3 temporal reasoning over causal chamber observations.
Metric: Accuracy (%). Source: huggingface.co. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | TimeClaw (deepseek-v4-pro) | 48.8 | |
| 2 | CAMEL (deepseek-chat) | 46.8 | |
| 3 | TimeCopilot (deepseek-v4-pro) | 46.8 | |
| 4 | AgentScope (deepseek-chat) | 46.4 | |
| 5 | MetaGPT (deepseek-chat) | 46.4 | |
| 6 | MetaGPT (gpt-4o) | 45.2 | |
| 7 | Single LLM (deepseek-chat) | 44.4 | |
| 8 | AgentScope (gpt-4o) | 44 | |
| 9 | CAMEL (gpt-4o) | 42 | |
| 10 | TimeCopilot (deepseek-chat) | 40 |
Interactive version: theaggregate.ai/benchmark?slug=temporalbench-causalchambers-t3 · How It Works · Data refreshed daily, snapshot 2026-10-20.