Tibo's Day 7 | GPT-6.1 Sol default context grows to 500k

GPT-6.1 Sol’s default context in Codex and Work increases from 272k to 500k tokens. Tibo’s Day 7 announcement says this aims to reduce compaction and retain more detail during long tasks.

A larger default; a 1M maximum

Item	The day’s announcement
Default context	272k → 500k tokens, described by the author as roughly 80% larger
User configuration	Users can still override the default; maximum model context remains 1M
Product scope	GPT-6.1 Sol in Codex and Work

Working across many files, referring repeatedly to earlier requirements or running long tasks can bring context compaction into view. A larger window leaves more room for that background. This explains the mechanism; CRW has not measured an improvement in task completion rates.

The subscription fee clarification

In a reply, Tibo says GPT-6.1 Sol has no long-context surcharge when using subscription usage. That answers the fee question for this model’s subscription path.

The context window determines how much information a task can hold; a plan allowance determines usage over time. The larger window is not an announcement of roughly 80% more available tasks. Normal input and output still consume usage.