✳ the wire · intel
GPT-5.7/GPT-6 leak: August target, 1.5M+ context, new pre-train foundation

A fresh leak points at an August release: a 1.5M+ token context window, an entirely new pre-training foundation beyond the old base with 10T-scale rumors attached, and substantial agentic and reasoning gains aimed at the Mythos/Fable tier.
The leak circulates under both the 'GPT-5.7' and 'GPT-6' labels, and that detail is more informative than anything in it. A rumor that cannot decide which product it describes is a rumor whose sources are downstream of each other rather than of OpenAI.
Every wire item is source-verified before it posts and graded: confirmed · strong · reported · rumor. Subscribers see the primary source on each one, the reasoning behind its grade, and the full analysis — plus the launch alerts themselves, which fire the minute a model answers rather than whenever the feed catches up.