Skip to content
  • Categories
  • Recent
  • Tags
  • Popular
  • World
  • Users
  • Groups
Skins
  • Light
  • Brite
  • Cerulean
  • Cosmo
  • Flatly
  • Journal
  • Litera
  • Lumen
  • Lux
  • Materia
  • Minty
  • Morph
  • Pulse
  • Sandstone
  • Simplex
  • Sketchy
  • Spacelab
  • United
  • Yeti
  • Zephyr
  • Dark
  • Cyborg
  • Darkly
  • Quartz
  • Slate
  • Solar
  • Superhero
  • Vapor

  • Default (No Skin)
  • No Skin
Collapse

NodeBB

  1. Home
  2. Blogs
  3. 关于 GPT-5.6 Sol 超长上下文阶梯计费的说明

关于 GPT-5.6 Sol 超长上下文阶梯计费的说明

Scheduled Pinned Locked Moved Blogs
openaigpt-5.6 sol超长上下文阶梯计费codex 配置
1 Posts 1 Posters 125 Views 1 Watching
  • Oldest to Newest
  • Newest to Oldest
  • Most Votes
Reply
  • Reply as topic
Log in to reply
This topic has been deleted. Only users with topic management privileges can see it.
  • A Online
    A Online
    admin
    wrote on last edited by
    #1

    <p>根据 OpenAI 最新计费规则,GPT-5.6 Sol 等模型对超长上下文请求采用阶梯价格。</p><p>当单次请求的上下文超过 272K tokens 时,输入、缓存读取以及输出可能按照更高阶梯价格计算。该规则来自 OpenAI 官方定价,并非 AI-ROUTER 临时加价或异常扣费。</p><p></p><p>以 GPT-5.6 Sol 标准价格为例:</p><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> 项目 272K 以内 超过 272K</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> ━━━━━━━━━━ ━━━━━━━━━━━━━━━━━━━ ━━━━━━━━━━━━━━━━━</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> 输入 $5 / 1M tokens $10 / 1M tokens</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> ────────── ─────────────────── ─────────────────</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> 缓存读取 $0.50 / 1M tokens $1 / 1M tokens</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> ────────── ─────────────────── ─────────────────</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> 输出 $30 / 1M tokens $45 / 1M tokens</code></pre><p></p><p>最终费用仍会根据具体分组、账号倍率及其他计费配置计算。</p><p></p><p>使用记录中出现 x2 标志,表示本次请求触发了长上下文阶梯计费。该标志主要表示输入/缓存读取进入更高阶梯,并不代表整条请求的所有费用简单翻倍;输出可能采用不同的阶梯倍率。</p><p></p><p><strong> ## 如何避免触发阶梯计费</strong></p><p>如果不需要 1M Token 上下文窗口,建议将 Codex 的上下文上限调整回 272K 以内。</p><p></p><p>打开:</p><p> ~/.codex/config.toml</p><p></p><p>将原先文章中的 1M 配置:</p><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> model = "gpt-5.6-sol"</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> model_context_window = 1000000</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> model_auto_compact_token_limit = 900000</code></pre><p> </p><p>调整为:</p><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> model = "gpt-5.6-sol"</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> model_context_window = 272000</code></pre><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code> model_auto_compact_token_limit = 250000</code></pre><p> </p><p>保存后重启 Codex,并新建会话。</p><p></p><p>如果只希望临时对单个会话生效,可以使用:</p><pre class="rounded-lg bg-slate-950 px-4 py-3 text-slate-100"><code>codex -m gpt-5.6-sol -c model_context_window=272000 -c model_auto_compact_token_limit=250000</code></pre><p> 同时建议:</p><p> - 及时开启自动压缩;</p><p> - 将超长任务拆分为多个会话;</p><p> - 减少一次性返回的大量工具输出;</p><p> - 定期总结并清理较早的对话内容。</p><p></p><p>如果确实需要 1M Token 上下文窗口,可以继续使用原配置,但超出 272K 后将按照 OpenAI 的高阶梯价格计费。</p><p></p><p>原始配置说明请参考:<a target="_blank" rel="noopener noreferrer nofollow" href="https://ai-router.dev/blog/post-g-677aa3e043912838-how-to-enable-a-1m-token-context-window-in-codex-for-gpt-5-6-sol">如何在 Codex 中为 GPT-5.6 Sol 启用 100 万 Token 上下文窗口 </a></p><p></p><p>感谢您的理解与支持。</p>

    1 Reply Last reply
    0

    Hello! It looks like you're interested in this conversation, but you don't have an account yet.

    Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.

    With your input, this post could be even better 💗

    Register Login
    Reply
    • Reply as topic
    Log in to reply
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes


    • Login

    • Don't have an account? Register

    • Login or register to search.
    Powered by NodeBB Contributors
    • First post
      Last post
    0
    • Categories
    • Recent
    • Tags
    • Popular
    • World
    • Users
    • Groups