ImplicitMemBench: leaderboard
Metric: Overall score: mean of procedural-memory accuracy, classical-conditioning accuracy (%) and the GPT-4o-mini-judged priming influence score (0 to 95 bands) on the 300 ImplicitMemBench items (learning, interference and test phases, first-attempt scoring, mean of three runs at temperature 0); higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek R1 | 65.3 |
| 2 | Qwen 3 32B | 64.13 |
| 3 | GPT-5 | 63 |
| 4 | Qwen 3 8B | 62.35 |
| 5 | O3 | 61.79 |
| 6 | O4 Mini (High) | 60.87 |
| 7 | GLM-4.5 | 58.59 |
| 8 | Gemini 2.5 Pro | 55.69 |
| 9 | Claude Opus 4.1 | 55.65 |
| 10 | Gemini 2.5 Flash | 55.43 |
| 11 | GPT-4o Mini | 50.88 |
| 12 | Qwen 2.5 72B Instruct | 50.78 |
| 13 | GPT-4o | 50.32 |
| 14 | Claude Sonnet 4 | 49.84 |
| 15 | Llama 3.3 70B Instruct | 49.44 |
Interactive version: theaggregate.ai/benchmark?slug=implicitmembench · How It Works · Data refreshed daily, snapshot 2026-10-07.