Sqz: Compress LLM context to save tokens and reduce costs

sqz vs rtk — Token Savings Benchmark: https://github.com/ojuschugh1/sqz/blob/main/docs/benchmark-v...