MAGMA
https://arxiv.org/abs/2601.03236
Multi-graph agentic memory architecture for AI agents.
At a glance
- Type
- Multi-graph agentic memory
- Tier
- T4
- Created
- 2026-01
- Latest release
- not applicable — not OSS
- License
- not applicable — not OSS
- GitHub
- not applicable — no GitHub repo
- Pricing
- not applicable — not commercial
- Funding
- not applicable — not commercial
Taxonomy
- storage
- graph
- retrieval
- graph-traversal
- persistence
- long-term
- update
- extraction
- unit
- episode
- governance
- inspectable
- conflict
- llm-arbitrate
When to use
Optimised for: not applicable - research paper
Anti-fit: not applicable - research paper
Pros & cons
Pros
Multi-agent shared memory framework.
Cons
Preprint; multi-agent coordination concerns at scale.
Claims & capabilities
Multi-graph memory for long-horizon reasoning across semantic/temporal/causal/entity graphs; policy-guided traversal; ACL 2026. Headline: 0.700 overall on LoCoMo, vs. Full-Context 0.481, A-MEM 0.580, MemoryOS 0.553, Nemori 0.590 (18.6%-45.5% relative gains); baselines: A-MEM, Nemori, MemoryOS, Full-Context; primary datasets: LoCoMo (~9K-token convs) and LongMemEval (100K+ token, 1,986 samples across 5 query categories).
Technical surface
- API surface
- not applicable — research paper
- Backend storage
- not applicable — research paper
- Deployment
- not applicable — not a deployable product
- Embedding model
- not applicable — research paper
- Multi-tenancy
- not applicable — research paper
- MCP
- not applicable — research paper, no deployed product
- A2A
- not applicable — research paper, no deployed product
- OpenTelemetry
- not applicable — research paper, no deployed product
Similar systems
Other recent method papers — theorized, no distinct product in the catalog, ranked by inbound references.
- Compressive Transformer T3
Maintains recent states in full resolution while compressing older memories with learned compression functions. DeepMind.
- MemGPT v2 / agent-tools T3
Already in catalog as the foundational MemGPT paper. Note: Letta is the productionised successor (cross-listed).
- Transformer-XL T3
Extends context through segment-level recurrence + caching of hidden states from prior segments. Foundational long-context architecture.
- Generative Agents T3
Park et al. — landmark agent-simulation paper. Reflection + memory stream + retrieval enable believable agent behavior.
- MemoryBank T3
Enhances LLMs with long-term memory. Early influential paper.
- Reflexion T3
Language agents with verbal reinforcement learning.
Related systems
References (4)
- A-MEM cites — S2 isInfluential citation
- LoCoMo cites — S2 isInfluential citation
- LongMemEval cites — S2 isInfluential citation
- Zep & Graphiti cites — S2 isInfluential citation