OpenAI 28 天活动第一周盘点：Astra 提速、额度重置
用户逐日复盘 OpenAI 活动前七天：Astra 与 Sol 提速约 50%、两次用量重置、Sol 上下文提到 500k，但整体觉得不够有趣。
Not gonna lie, I'm pretty disappointed with the whole 28 days of OpenAI features. Today is day 8, the second week of this "event". I hope they start bringing us more fun stuff! Here's what we got in the first week: Day 1: Faster models by default. GPT-6 Astra and GPT-6.1 Sol became approximately 50% faster. Day 2: Four improvements, plus a reset. Free Auto-review: “Approve for me” stopped consuming your plan’s usage allowance. Simpler API tiers: Five paid tiers became Build, Launch, and Grow. Qualifying for the highest tier dropped from $1,000 to $500 in cumulative API payments. Meetings plugin: Meeting notes, personalized summaries, and next steps saved into ChatGPT Space. Decisions API public beta: Apps can select models, tools, or actions in near real time. After a community vote, Tibo also announced a full usage reset, while keeping the improvements. Day 3: GPT-6 in Chat, a banked reset, and Codex Cloud. Day 4: Instant steering and Sol Ultrafast. Codex became quicker to react when you redirect an ongoing task. GPT-6.1 Sol Ultrafast launched in the API, Codex, and ChatGPT Work, advertised at up to 8× the speed of Sol Standard. Day 5: Next-message predictions and mobile dots. Composer predictions suggest your next message in the desktop app, with prediction generation included for Pro users without consuming usage. Users also gained the ability to create and message their dot entirely from the ChatGPT mobile app. Day 6: Global usage reset. Day 7: Larger default context without a long-context surcharge. GPT-6.1 Sol’s default context in Codex and Work increased from 272k to 500k tokens, reducing the need to compact long conversations. Tibo also confirmed no long-context fees when drawing usage from a subscription.