Managing Procedural Memory in LLM Agents: leaderboard
Metric: size (self-reported). Source: benchmarklist.com. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 9B | 3 |
| 2 | Gemma 4 E4B | 3 |
| 3 | GPT-OSS-20B | 2 |
| 4 | Gemma 4 31B | 2 |
| 5 | Gemma 4 26B A4B | 2 |
| 6 | Qwen 3.5 35B A3B | 2 |
| 7 | GPT-5.4 Mini | 2 |
| 8 | GPT-5.4 | 1 |
| 9 | GPT-OSS-120B | 1 |
| 10 | Qwen 3.5 122B A10B | 1 |
| 11 | DeepSeek V4 Flash | 1 |
Interactive version: theaggregate.ai/benchmark?slug=managing-procedural-memory-in-llm-agents · How It Works · Data refreshed daily, snapshot 2026-09-05.