Grounding Awareness - Mem0 - Sensitivity: leaderboard
Grounding Awareness - Mem0 - Sensitivity. Paired correctness, 800 individual trajectories per agent-memory configuration. GPT-5.4-mini, Qwen3.8-Flash, Gemini-3.1-Flash-Lite and paper-unspecified DeepSeek-Flash backbone; no inferred DeepSeek version. Metric: Sensitivity (0–1). Primary published table; source configurations remain distinct.
Metric: Sensitivity (0–1). Source: arxiv.org. Saturation forecast: Around 2029. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.8 Flash | 0.13 | #45 |
| 2 | GPT-5.4 Mini | 0.09 | #207 |
| 3 | Gemini 3.1 Flash Lite | 0.03 | #187 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=grounding-awareness-mem0-sensitivity · How It Works · Data refreshed daily, snapshot 2026-10-16.