Leaked signals and official updates suggest a busy June for AI coding models. Anthropic's Claude Mythos (a frontier model with strong coding and cybersecurity capabilities) has already scanned 1,000+ open-source projects and may be heading to Claude Code and a new Claude Security dashboard. Claude Opus 4.8 is also rumored, spotted on Google Vertex, and could serve as a more accessible flagship upgrade for regular users. On the OpenAI side, GPT-5.6 labels have appeared in Codex routing logs and internal model tags (Iris Alpha, Ember Alpha, Beacon Alpha), though nothing is officially confirmed. OpenAI did confirm an internal reasoning model solved an 80-year-old Erdős math conjecture, hinting at stronger reasoning capabilities that could eventually reach Codex.

9m watch time

Sort: