Here is how to enable a 1M-token context window in Codex for GPT-5.6 Sol. Even though we have tuned the context limit in Codex to be set optimally when it comes to performance and cost, this is a common ask, so here it is documented. A larger context window lets Codex retain m…↗

no, there won't be a pi release for fable today. do this: "hey pi, here are the stats for anthropic's fable re: context window/cost. add it to models.json" enjoy. just don't do bio or cyber!↗

Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens -…↗

Asking Claude Code to make an HTML visualizing its current context window is a pretty wild way to learn about how context window works↗

I think 200k context window w/ good compaction is the sweet spot. Models claiming to have larger are rarely better at rembering important stuff from my experience↗

x.com/i/article/2044…↗

Harness, Memory, Context Fragments, & the Bitter Lesson this is a work in progress mental dump on interesting intersections between how we use and design a harness, implications for memory being accumulated over long timescales, and the search bitter lesson we can’t escape this…↗

Underrated claude code trick - use pm2 to run the dev server Makes it super easy for claude code to search logs while efficiently managing the context window↗

Today we’re shipping 3 more updates for Claude Code: - Claude Code on Android - Hotkey model switcher - Context window info in status lines↗

Today, code-supernova becomes code-supernova-1-million. This new stealth model is a supercharged version of code-supernova that comes with a 1M context window. Available for a limited time to individual users for free. Excited to hear what you build with it!↗