Tibo@thsottiaux
Day 7/ Summary For GPT-6.1 Sol in Codex and Work, the default context grows from 272k to 500k, about an 80% increase. The larger working context should reduce compaction and retain more detail during long tasks. Users can still override the setting up to the model’s 1M maximum, and further default increases are planned.
Show Chinese translation
第 7 天 / 内容摘要 在 Codex 与 Work 中使用 GPT-6.1 Sol 时，默认上下文从 272k 扩大到 500k，约增加 80%。这有助于减少上下文压缩，并保留长任务的细节。用户仍可自行调高或调低配置，该模型最大上下文为 1M；Tibo 表示未来还计划继续增加默认长度。