How long does your prompt processing actually take when resuming a long session?

<!-- SC_OFF --><div class="md"><p>How do you guys deal with long context, for example while coding, when you’re going back and forth for adjustments or fixing some errors and

<!-- SC_OFF --><div class="md"><p>I actually make some tests with lm studio and Opencode with the new Gemma 4 26b model.</p> <p>The results are really impressive especially on

<!-- SC_OFF --><div class="md"><p>I run pretty large models (DS4 Flash, Qwen3.8 Flash, ...) on a machine with pretty slow prompt processing (M2 Ultra).</p> <p>The first prompt

<!-- SC_OFF --><div class="md"><p>I organize everything by topic (work, study, documents, buying, recipes, etc.), and each topic stays in one long chat. Over time, those chats

<!-- SC_OFF --><div class="md"><p>im using chatgpt for studying, so i ask alot and prefer to have long conversations while it is explaning to me and i usually mention something i

<!-- SC_OFF --><div class="md"><p>Mainly, i'd like to understand how ChatGPT handles context compaction in long and complex conversations. How efficient is the retrieval of key