GPT-5.6 leaks from Codex logs — 1.5M token context window would be largest among frontier models, up from ~1.05M on GPT-5.5
GPT-5.6 references have surfaced in Codex agent logs ahead of any OpenAI announcement. The leaked context window of approximately 1.5M tokens would establish GPT-5.6 as the largest-context frontier model, up from ~1.05M on GPT-5.5. The context-window inflation continues a 12-month trend that has erased Gemini's prior structural lead in long-context workloads.
The substantive piece is the context-window competitive flattening. Eighteen months ago Gemini's 2M-token context window was a structural differentiator that justified premium pricing for long-document workloads. The Anthropic Fable-5 and rumored GPT-5.6 windows close that gap; Llama 4 Scout's 10M context on single H100 overtakes it entirely on the open-source side. Context window has decommissioned as a moat for any single vendor; procurement evaluations should now focus on retrieval-quality-at-length metrics rather than raw window size.
The OpenAI strategic read is that GPT-5.6 is positioned to neutralize the long-context-leadership narrative ahead of any IPO commentary about competitive position. The Codex-log surface is consistent with internal dogfooding ahead of public availability; the historical pattern suggests a release within 4-8 weeks of leak surfacing. Procurement teams running competitive Claude-vs-GPT-vs-Gemini evals on long-context workloads through Q3 2026 should price the GPT-5.6 release into their evaluation timelines.
Essa Mamdani — June 2026 AI Model Flood: GPT-5.6, Gemini 3.5 Pro & Claude 4.8 → · Centerbit — AI Rumors June 2026: GPT-5.6, Gemini 3.5 Pro, Claude Mythos → · Stob.AI — Best AI Model 2026: GPT-5.5 vs Claude 4.8 vs Gemini 3.5 vs Llama 4 [Tested] →