1 Million Context to enable professional workloads

A slightly contrarian observation from measuring long-running Codex workloads:

I am no longer convinced that a larger context window is always the right metric for professional agent workloads.

In our audit, the expensive failure mode was not “the model cannot fit enough information.” It was that completed execution history remained economically active and was repeatedly processed during later orchestration.

We measured individual MAIN turns above 200k input tokens in the old workflow, including very large turns whose semantic purpose was essentially checking external progress.

A 1M window certainly postpones compaction, but it may also allow much more completed history to remain active before the system is forced to deal with it.

I would therefore love to see benchmarks report not only maximum context supported, but peak active context required per successfully completed task.

Bigger RAM and better memory management solve different problems.

While a 1M token window could help I think most people do not manage their context properly. I used to run out of context a year ago when I was just getting started with agentic coding but it has not happened to me in months.

Important thing: after exceeding 272k input tokens, GPT-5.6 has a different cost threshold - the API documentation states 2x input cost and 1.5x output cost for the entire request.

Personally, I think the context window should be set based on the comprehension level. If I want maximum performance, I’ll set it to ultra and have 1 million (which is logical, as there will be many logical variants to remember). If I’m running a small project and set the comprehension level to a lower level, I’ll use a smaller context window.

This is just a suggestion.