OpenAI 的 Astra 模型用户，尤其是在 Plus 订阅用户中，报告称 token 消耗量极高，很快就会耗尽他们的使用限额。许多人认为 Astra 的快速消耗存在问题，质疑其与 GPT-5.6 Sol 等其他模型相比的价值，后者被认为在复杂任务上更具成本效益和可靠性。一些用户正在开发变通方法，例如先使用 Astra 进行初步规划，然后切换到其他模型进行执行，以管理其使用。 AI
影响 Astra 的高 token 消耗量可能迫使用户限制其使用或寻求替代模型，从而影响工作流程效率。
排序理由 用户对特定产品功能（Astra 模型的使用量）的讨论和抱怨。
AI 生成摘要 · Google Gemini · 来自 5 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
<!-- SC_OFF --><div class="md"><p>Never used to run into usage issues on ChatGPT (that annoyance was reserved for Claude) but Astra just burns through tokens. Two prompts concerning my codebase rips through my 5 hr cap. Any tips on reducing usage? </p> </div><!-- SC_ON -->   …
<!-- SC_OFF --><div class="md"><p>Hey! I honestly don't really understand the point of the new Astra model. I've seen tons of Plus users complaining about how insanely fast it burns through usage, and honestly, they're right to complain. I don't understand why Sam Altman decided …
<!-- SC_OFF --><div class="md"><p>I've been doing some not crazy stuff on Opencode using Astra on low... its BLASTING through usage. Is this normal?? Maybe i was dumb but i assumed since Astra and GPT 5.6 Sol weren't that far off from eachother, it wouldn't be huge token usage at…
<!-- SC_OFF --><div class="md"><p>If you’re on Plus, I really wouldn’t use ASTRA for the whole coding task. It can burn through the 5 hour limit insanely fast, sometimes in just 1 prompt and 10 minutes.<br /> What I’ve been doing instead:<br /> Use <strong>ASTRA first for plannin…
<!-- SC_OFF --><div class="md"><p>Was excited to use my 3 banked resets to get astra to bang out a bunch of tasks I had lined up but streets are saying it’ll casually blow through the 5 hour in like 4 turns and now I’m too afraid to try it out lol.</p> </div><!-- SC_ON -->   …