inference 2 Half the Cache, Twice the Context — measuring what KV quantization costs at long range Sep 26, 2026 How to Self-Host Open-Weight LLMs: A Practical Guide Jul 31, 2026