Cognee
Community-VerifiedMemory management layer that builds dynamic knowledge graphs from conversations and documents. Combines semantic chunking with graph-based retrieval for contextual recall.
Scores from 0–100. Higher is better. LLM Baseline (no memory system) scores 57.6%. How we calculate this →
TrackKnowledge Graph
Track IndexNo results yet
Benchmark Results
| Benchmark | Score | Status | Receipt |
|---|---|---|---|
| Knowledge Retrieval | Pending | Pending | -- |
| Knowledge Scale | Pending | Pending | -- |
| Truth Arbitration | Pending | Pending | -- |
| Budget Curves | Pending | Pending | -- |
| Reliability | Pending | Pending | -- |
| Other Benchmarks | |||
| LongMemEval | Not applicable — outside Knowledge Graph track | ||
| LoCoMo | Not applicable — outside Knowledge Graph track | ||
| Memory Poisoning | Not applicable — outside Knowledge Graph track | ||
Relative Performance vs All Benchmarked Systems
vs 16 scored systemsEach dot is a system. Amber dot is Cognee. Amber line = LLM Baseline (no memory).
Overall0.00th percentile
No memory: 57.6%gbrain
Recall0.00th percentile
No memory: 57.6%gbrain
Temporal0.00th percentile
No memory: 57.6%gbrain
Reasoning0.00th percentile
No memory: 57.6%gbrain
Bench'd Memory Index
The BMI combines accuracy (70%) and efficiency (30%) into a single production-weighted score. Formula is public and versioned.
0.0
/ 100
#7 of 8 systemsTop 87%
Accuracy (70%)0.0
Efficiency (30%)--
Per-Capability Score Matrix
| Dimension | LongMemEval | Reliability |
|---|---|---|
| Recall | 50.0 | -- |
| Temporal | 0.0 | -- |
| Reasoning | 0.0 | -- |
| Hallucination | -- | 0.0 |
| Stale Memory | -- | 0.0 |
| Entity Confusion | -- | 0.0 |
| Deletion | -- | 0.0 |
| Overall | 20.0 | 0.0 |
Per-Benchmark Breakdown
| Benchmark | Verified | Nuance |
|---|---|---|
| LongMemEval | 0.0 | 0.0 |
| LongMemEval | 79.6 | 73.1 |
| LoCoMo | 78.3 | 72.0 |
Performance Over Time — LongMemEval
2026-05-11 to 2026-05-13Cognee20.0
Most often compared with
Add badge to your README
Show your Bench'd score on your GitHub repo.
Markdown
[](https://benchd.ai/system/cognee)
HTML
<a href="https://benchd.ai/system/cognee"><img src="https://img.shields.io/badge/Bench'd_BMI-0.0-D9982B?style=flat" alt="Bench'd Verified: 0.0 BMI" /></a>