MemoryStackBench Leaderboard

Quantitative memory safety check scores from local and CI benchmark runs.

Overall Check Scores

100%0%100%agno_memory100%aws_bedrock_ag100%cognee100%crewai_memory100%google_adk_mem100%hindsight100%langgraph100%langmem100%llamaindex_mem100%supermemory100%zep94%letta91%autogen_mem0me91%mem091%openai_agents_79%tencentdb-agen73%graphiti

Ranked Runs

#RunTargetFrameworkSuiteChecksScenariosFailures
1agno-memory-localagno_memory_pinnedagno_memoryseven_sins_v0_133/33 (100%)5/5 (100%)0
2aws-agentcore-memory-localaws_bedrock_agentcore_memory_pinnedaws_bedrock_agentcore_memoryseven_sins_v0_133/33 (100%)5/5 (100%)0
3cognee-localcognee_local_pinnedcogneeseven_sins_v0_133/33 (100%)5/5 (100%)0
4crewai-memory-localcrewai_memory_pinnedcrewai_memoryseven_sins_v0_133/33 (100%)5/5 (100%)0
5google-adk-memory-bank-localgoogle_adk_memory_bank_pinnedgoogle_adk_memory_bankseven_sins_v0_133/33 (100%)5/5 (100%)0
6hindsight-localhindsight_pinnedhindsightseven_sins_v0_133/33 (100%)5/5 (100%)0
7langgraph-locallanggraph_persistence_pinnedlanggraphseven_sins_v0_133/33 (100%)5/5 (100%)0
8langmem-locallangmem_pinnedlangmemseven_sins_v0_133/33 (100%)5/5 (100%)0
9llamaindex-memory-localllamaindex_memory_pinnedllamaindex_memoryseven_sins_v0_133/33 (100%)5/5 (100%)0
10supermemory-api-localsupermemory_api_pinnedsupermemoryseven_sins_v0_133/33 (100%)5/5 (100%)0
11zep-cloud-localzep_cloud_or_self_hosted_pinnedzepseven_sins_v0_133/33 (100%)5/5 (100%)0
12letta-localletta_self_hosted_pinnedlettaseven_sins_v0_131/33 (94%)3/5 (60%)2
13autogen-mem0-localautogen_mem0memory_pinnedautogen_mem0memoryseven_sins_v0_130/33 (91%)3/5 (60%)3
14mem0-localmem0_open_source_pinnedmem0seven_sins_v0_130/33 (91%)3/5 (60%)3
15openai-agents-sessions-localopenai_agents_sdk_sessions_pinnedopenai_agents_sdk_sessionsseven_sins_v0_130/33 (91%)3/5 (60%)3
16tencentdb-agent-memory-localtencentdb_agent_memory_pinnedtencentdb-agent-memoryseven_sins_v0_126/33 (79%)2/5 (40%)7
17graphiti-neo4j-localgraphiti_neo4j_pinnedgraphitiseven_sins_v0_124/33 (73%)1/5 (20%)9

Category Breakdown

agno_memory_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

aws_bedrock_agentcore_memory_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

cognee_local_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

crewai_memory_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

google_adk_memory_bank_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

hindsight_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

langgraph_persistence_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

langmem_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

llamaindex_memory_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

supermemory_api_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

zep_cloud_or_self_hosted_pinned

Checks 33/33 (100%) · Scenarios 5/5 (100%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance100%
write_correctness100%

letta_self_hosted_pinned

Checks 31/33 (94%) · Scenarios 3/5 (60%)
deletion_behavior100%
retrieval_correctness83%
temporal_update_handling100%
untrusted_source_resistance90%
write_correctness100%

autogen_mem0memory_pinned

Checks 30/33 (91%) · Scenarios 3/5 (60%)
deletion_behavior60%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance90%
write_correctness100%

mem0_open_source_pinned

Checks 30/33 (91%) · Scenarios 3/5 (60%)
deletion_behavior60%
retrieval_correctness100%
temporal_update_handling100%
untrusted_source_resistance90%
write_correctness100%

openai_agents_sdk_sessions_pinned

Checks 30/33 (91%) · Scenarios 3/5 (60%)
deletion_behavior100%
retrieval_correctness100%
temporal_update_handling86%
untrusted_source_resistance80%
write_correctness100%

tencentdb_agent_memory_pinned

Checks 26/33 (79%) · Scenarios 2/5 (40%)
deletion_behavior40%
retrieval_correctness100%
temporal_update_handling86%
untrusted_source_resistance70%
write_correctness100%

graphiti_neo4j_pinned

Checks 24/33 (73%) · Scenarios 1/5 (20%)
deletion_behavior100%
retrieval_correctness50%
temporal_update_handling71%
untrusted_source_resistance70%
write_correctness80%

Target Coverage

TargetFrameworkStatusRuntimeScoreRunBlockers / Notes
agno_memory_pinnedagno_memoryimplemented_store_harnesslocal_python100%agno-memory-local
autogen_mem0memory_pinnedautogen_mem0memoryimplementedlocal_python91%autogen-mem0-local
aws_bedrock_agentcore_memory_pinnedaws_bedrock_agentcore_memoryimplemented_store_harnessexternal_service100%aws-agentcore-memory-local
cognee_local_pinnedcogneeimplemented_store_harnesslocal_python100%cognee-local
crewai_memory_pinnedcrewai_memoryimplemented_store_harnesslinux_container100%crewai-memory-local
google_adk_memory_bank_pinnedgoogle_adk_memory_bankimplemented_store_harnessexternal_service100%google-adk-memory-bank-local
graphiti_neo4j_pinnedgraphitiimplementeddocker_compose73%graphiti-neo4j-local
hindsight_pinnedhindsightimplemented_store_harnessdocker_or_external_service100%hindsight-local
langgraph_persistence_pinnedlanggraphimplemented_store_harnesslocal_python100%langgraph-local
langmem_pinnedlangmemimplemented_store_harnesslocal_python100%langmem-local
letta_self_hosted_pinnedlettaimplementeddocker_compose94%letta-local
llamaindex_memory_pinnedllamaindex_memoryimplemented_store_harnesslocal_python100%llamaindex-memory-local
mem0_open_source_pinnedmem0implementedlocal_python91%mem0-local
openai_agents_sdk_sessions_pinnedopenai_agents_sdk_sessionsimplementedlocal_python91%openai-agents-sessions-local
supermemory_api_pinnedsupermemoryimplementedexternal_service100%supermemory-api-local
tencentdb_agent_memory_pinnedtencentdb-agent-memoryimplementedlocal_node_gateway79%tencentdb-agent-memory-local
zep_cloud_or_self_hosted_pinnedzepimplemented_store_harnessexternal_service100%zep-cloud-local