Recent multiple degraded-performance incidents in OpenAI Codex, including an eight-hour GPT-6.1 Sol disruption on October 13 affecting the API, VS Code extension, CLI, and web interface plus a follow-up event the next day, have reinforced trader expectations of further outages before year-end. Persistent GitHub reports of capacity errors, server overloads, and 429/503 responses even outside official incidents point to underlying scaling challenges with high-demand large language models. These reliability patterns, combined with ongoing heavy usage of Codex tools and no announced architectural fixes, underpin the 71% market-implied odds for another 2026 outage while leaving room for mitigation if capacity management improves.