Evaluating Mem0, Letta (MemGPT), Zep + Graphiti, Qdrant, Milvus, LightRAG, and pgvector on long-horizon recall and tenant memory isolation.
Task Performance (30%) + Proactivity (20%) + Memory (15%) + Privacy (15%) + Independence (10%) + Real Usage (10%) [Social Volume = 0%].
2. Ranker Score Formula: Methodology Transparency (30%) + Coverage (25%) + Freshness (20%) + Independence (15%) + Reproducibility (10%).
3. Category Leaders: Agentic Models #1 Gemini 4 Argon High (93.1), #2 DeepSeek-V4 (92.9), #3 Grok 4.7 (92.3), #4 Llama Spark 1.3 Max (91.9); Workflow Builders #1 n8n AI Agent (85.3); Personal Agents #1 Grok Bot (88.5), #2 Scout (88.4), #3 Muse (86.9), while heavily hyped social-pump entries like Dots (#5 at 73.4) and closed chat wrappers (Claude Sonnet 5.5 at 79.6, GPT-6 Astra Max at 78.9) rank lower due to unverified autonomy claims, rate-limit throttling, and vendor lock-in.
| Rank | Agent & Direct Link | Category | Score | Tier Badge | Hype Signal | Empirical Audit Summary |
|---|---|---|---|---|---|---|
| #4 |
Qdrant • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 89.0 | 85+ ✅ Excellent | ✅ Verified | High-performance Rust vector database with payload filtering and multi-tenant memory isolation. |
| #2 |
Letta (UC Berkeley Spinout) • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 88.9 | 85+ ✅ Excellent | ✅ Verified | OS-inspired hierarchical virtual context management (Core Memory vs Archival Recall) for perpetual agents. |
| #5 |
Zilliz / LF AI • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 88.3 | 85+ ✅ Excellent | ✅ Verified | Linux Foundation open-source GPU-accelerated vector database for enterprise agent swarms. |
| #1 |
Mem0.ai • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 88.1 | 85+ ✅ Excellent | ✅ Verified | Hybrid vector + graph + key-value episodic memory layer reducing multi-session agent token cost by 80%. |
| #11 |
PostgreSQL Community • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 87.8 | 85+ ✅ Excellent | ✅ Verified | ACID-compliant transactional agent memory directly inside existing PostgreSQL clusters. |
| #7 |
HKU Data Intelligence Lab • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 87.3 | 85+ ✅ Excellent | ✅ Verified | Dual-level retrieval graph RAG achieving 10x lower token overhead than Microsoft GraphRAG. |
| #3 |
Zep AI • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 87.2 | 85+ ✅ Excellent | ✅ Verified | Bi-temporal knowledge graph tracking how user and enterprise facts evolve over time with sub-50ms retrieval. |
| #8 |
LanceDB • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 86.6 | 85+ ✅ Excellent | ✅ Verified | Embedded columnar Arrow-native vector store for local on-device agent episodic memory. |
| #15 |
InfiniFlow • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 85.6 | 85+ ✅ Excellent | ✅ Verified | Open-source complex table, PDF, and financial report chunking with grounded citations. |
| #6 |
Cognee (topoteretes) • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 85.5 | 85+ ✅ Excellent | ✅ Verified | Modular ECL (Extract, Cognify, Load) relational + graph memory pipeline for Python agents. |
| #14 |
LlamaIndex • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 85.5 | 85+ ✅ Excellent | ✅ Verified | Event-driven document parsing (LlamaParse) and agentic retrieval orchestration. |
| #9 |
Chroma • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 85.2 | 85+ ✅ Excellent | ✅ Verified | Developer-friendly Apache 2.0 local and client-server context store. |
| #10 |
Weaviate • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 85.2 | 85+ ✅ Excellent | ✅ Verified | Cloud-native modular vector database with built-in multi-tenancy. |
| #13 |
FalkorDB • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 84.2 | 75-84 ✅ Very Good | ✅ Verified | Sparse-matrix linear algebra property graph designed specifically for low-latency LLM agents. |
| #12 |
Neo4j • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 83.8 | 75-84 ✅ Very Good | ✅ Verified | Deterministic multi-hop entity relationship traversal for enterprise compliance agents. |
| #16 |
Redis • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 83.2 | 75-84 ✅ Very Good | ✅ Verified | Sub-millisecond in-memory sliding window and semantic cache for high-throughput agents. |
| #17 |
Pinecone • Last verified: 2026-10-27 • Report Outdated Link
|
memory-context | 76.1 | 75-84 ✅ Very Good | 🔥 High social | Reliable managed SaaS vector search, but lacks open-source self-hosting weights/binaries. |