August 19, 2026· ~11 min readDeepSeek V4 Flash 0731: Initial Experiences on 2x DGX Spark ## Overview DeepSeek AI released **DeepSeek-V4-Flash-0731**, an updated checkpoint of their mixture-of-experts (MoE) model that power their flaghshi
July 30, 2026· ~6 min readApplying the CaRE Protocol to Agent Benchmarks: Compute-Aware Evaluation for Skill and Memory Harnesses > **Teaser:** CaRE — the compute-aware remasking evaluation protocol that reversed published MDLM rankings — has been applied to our two open-source