ZeroMemory Holds 100% Recall on LongMemEval V2 — Now With Open Llama 70B Instead of GPT-4o
We re-ran the LongMemEval benchmark against ZeroMemory with Llama 3.3 70B instead of GPT-4o. Result: 100% Recall@1, zero errors, and 94% QA accuracy — all
We re-ran the LongMemEval benchmark against ZeroMemory with Llama 3.3 70B instead of GPT-4o. Result: 100% Recall@1, zero errors, and 94% QA accuracy — all