List your company

Memory Capability Evaluation on BEAM 100K tokens

5Abstention RateMemoryOS
222.2542.562.75Oct 31, 2025Dec 4, 2025Jan 7, 2026Feb 11, 2026Mar 17, 2026Apr 20, 2026May 25, 2026

Evaluation results

MethodLinks
MemoryOSJudge LLM=GPT-4.1-mini2026.05510.512.736.452.640.925.254.522.426.228.64
LightMemJudge LLM=GPT-4.1-mini2026.051015.418.150.845.641.63059.828.524.832.46
SimpleMemJudge LLM=GPT-4.1-mini2026.0512.520.625.160.7605538.168.130.73540.68
Llama Maverick (Vanilla)Model=Llama Maverick, Setup=Vanilla2025.10202.5195141.23015.2456.51024
Mem0Judge LLM=GPT-4.1-mini2026.052018.422.658.457.845.133.352.530.931.537.05
NEMORIJudge LLM=GPT-4.1-mini2026.052518.920.463.564.552.132.46431.73240.45
Qwen 2.5 (Vanilla)Model=Qwen 2.5, Setup=Vanilla2025.10303.119.242.54043.722.255.412.811.228
MEMIRJudge LLM=GPT-4.1-mini2026.0537.532.328.669.865.658.443.771.436.838.548.26
LIGHTModel=Qwen 2.5, Setup=LIGHT2025.1047.53.720.547.936.236.228.156.623.211.231.1
GPT-4.1-nano (Vanilla)Model=GPT-4.1-nano, Setup=Vanilla2025.1047.51.218.127.342.527.517.843.72.811.223.9
LIGHTModel=GPT-4.1-nano, Setup=LIGHT2025.1057.53.117.753.84037.536.562.520.216.234.5
LIGHTModel=Llama Maverick, Setup=LIGHT2025.10603.116.651.841.24535.362.523.818.735.8
Qwen 2.5 (RAG)Model=Qwen 2.5, Setup=RAG2025.10652.520.133.837.527.520.337.97.416.226.9
LIGHTModel=Gemini 2 Flash, Setup=LIGHT2025.1067.51.816.646.436.23022.546.213.912.529.4
Llama Maverick (RAG)Model=Llama Maverick, Setup=RAG2025.10803.116.239.237.53522.551.211.127.532.3
Gemini 2 Flash (Vanilla)Model=Gemini 2 Flash, Setup=Vanilla2025.10800.618.133.327.512.520301.818.724.2
Gemini 2 Flash (RAG)Model=Gemini 2 Flash, Setup=RAG2025.1080519.134.128.732.514.841.69.31528
GPT-4.1-nano (RAG)Model=GPT-4.1-nano, Setup=RAG2025.10801.816.936.23537.526.3558.312.530.9