TriggerBench: leaderboard
While Large Language Models (LLMs) are increasingly deployed in long interactions, existing evaluations focus predominantly on retrospective memory (RM) via explicit queries. Prospective memory (PM), the critical ability to spontaneously recall and act on latent constraints without direct prompts, remains largely unevaluated. We introduce TriggerBench, a comprehensive PM benchmark spanning five di
Metric: Slot Match (Positive Clean) (self-reported). Source: benchmarklist.com. Saturation forecast: Around April 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 90.88 | #110 |
| 2 | Gemma 3 27B (IT) | 79 | #523 |
| 3 | GPT-4.1 | 72.37 | #248 |
| 4 | GPT-4o | 60.38 | #343 |
| 5 | Qwen 3 32B | 49.32 | #443 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=triggerbench · How It Works · Data refreshed daily, snapshot 2026-10-20.