Quick links
Loading the latest distributed benchmark data…
Preview data uses the same DeepSWE benchmark and the existing cached interface. IQ and duration are directly comparable; cost is a reference/equivalent cost.
Inputs above 272K tokens do not incur additional long-context pricing multipliers. Codex also does not charge for cache writes.
Astra's context limit in Codex is 872,000 tokens. At 90%, the automatic compaction threshold is 784,800 tokens.
872,000 × 90% = 784,800
Pricing source：OpenAI official rate card ↗
Copy a task's deep link and give it to the other task. The link acts like a precise task address, so each side can identify the intended session before review, handoff, or coordinated work begins.
How: Open the target task's menu, choose Copy → Copy deep link, then paste the link into the other Codex task and state what each side should do.
Why it matters: The receiving task gets an exact, referable target instead of an ambiguous phrase such as “the other window.” This is useful for mutual review, progress handoff, and having one task consolidate another task's findings.
Boundary: A deep link establishes identity and navigation; it does not automatically synchronize permissions, working directories, or conversation context. Keep responsibilities explicit, and isolate simultaneous code changes with separate worktrees or an equivalent workflow.
Open Codex Settings → Configuration → Model features → Available reasoning efforts, then select Max. It will appear in model controls for supported models.
For a quick Codex quota conversion, 250 credits correspond to 10 US dollars of usage quota.
Automatic recommendations by task scenario based on Composite Intelligence.
Current recommendations focus on models available through a Codex subscription; DeepSeek uses separate API billing and will be evaluated and recommended separately. LIVEOnly sustained IQ declines versus the 24/48-hour averages.
LIVELoading update time
Community-tested data · every contribution makes the results more accurate. Contribute now →
Select one or more configurations to compare their history.
No configurations selected.
API price estimate (Standard, short context; input ≤272K): Luna $0.20/1M × 272K ≈ $0.0544; Sol $4/1M × 272K ≈ $1.088; output tokens billed by actual usage. Luna pricing · Sol pricing · As of 2026-08-25
Up to 20 runs per desktop view; swipe or use the arrows for older history.
Method: Sol, Terra, and Luna use low reasoning effort with the same count-from-1-to-1024 output task. Standard and Fast each run three independent samples and use the arithmetic mean; order alternates, up to three model pairs run concurrently, and each pair still runs Standard/Fast serially. Every sample uses a fresh app-server and ephemeral thread plus a random nonce. Complete outputs with cached_input_tokens = 0 are preferred. For this run, slow samples and complete cache-mismatched pairs are retained as real same-day experience; incomplete outputs are rejected and rerun. TTFT is delay to the first visible token, TPS is sustained token generation speed, and E2E is total time from submission to the complete result. Headline acceleration values are calculated as mean Standard divided by mean Fast; per-model E2E acceleration uses that model's Standard divided by Fast.
Events are assigned to months in Beijing time. Direct resets count independent landed actions; an announcement, confirmation, and propagation observation for the same action count once. Monthly reset-card totals include only independently confirmed delivered/completed batches; announcements or pending issuance do not count. The 500k figure is a recipient count, not 500k events. Included are the June 12 initial saved-reset rollout, the June 18 and June 30 additional banked resets, the July 2 AIE World's Fair worldwide grant, the July 13 500k-user compensation, the July 14 all-user grant, and the Aug 22 20M milestone grant; September is currently retained at one card pending further delivery confirmation. The referral campaign's actual successful delivery total was not public and is not estimated. Zero means no matching record in the current evidence set, not proof that none occurred.
What time is safest for distributing reset cards?
Of 37 verifiable resets in Codex Radar history, 23 occurred from 00:00 to 08:59 Beijing time; 08:00–08:59 is the peak hour with 7 resets. The site treats 16:00–22:00 as a relatively safer card-distribution window; this is a historical pattern, not a guarantee.
PreviewAstra was probably our biggest competitive advantage while it wasn’t generally available. Since we’ve had it our productivity jumped so much that we shifted some of our plans 6 months ahead and will ship them at DevDay…
OriginalAstra was probably our biggest competitive advantage while it wasn’t generally available.
Since we’ve had it our productivity jumped so much that we shifted some of our plans 6 months ahead and will ship them at DevDay instead of mid next year.
Chinese translationAstra 在还没有普遍开放的时候，可能是我们最大的竞争优势。 自从我们有了它，我们的生产力提升了很多，以至于我们把一些计划提前了 6 个月，会在 DevDay 发布，而不是明年年中。
Phrasesgenerally available · will ship them at DevDay
Model context reading · Belgian Francophone phrasing lens“Ship” means release or deliver product work here. The post is about productivity and launch timing, not Codex quota, rate limits, or resets.
Preview@lydiahallie Wow, huge, wonder why!
Original@lydiahallie Wow, huge, wonder why!
Chinese translation@lydiahallie 哇，很大，想知道为什么！
Model context reading · Belgian Francophone phrasing lensThis is a short reaction to someone else’s Claude limit-reset announcement. It may be joking or competitive commentary, but it does not itself promise a Codex reset.
Preview@theo Let the potato rest a little
Original@theo Let the potato rest a little
Chinese translation@theo 让土豆稍微歇一会儿吧
Phrasesrest
Model context reading · Belgian Francophone phrasing lens“Rest” is playful wording here, not “reset.” In context it is a casual reply about GPT-6 Sol excitement, not a quota-window signal.
Preview@DavidOndrej1 You’ve just revealed our strategy. See you at DevDay
Original@DavidOndrej1 You’ve just revealed our strategy. See you at DevDay
Chinese translation@DavidOndrej1 你刚刚透露了我们的战略。DevDay 见。
Model context reading · Belgian Francophone phrasing lensA playful event-facing reply. “strategy” and “DevDay” do not indicate usage, limits, resets, or banked reset credits.
PreviewAnd the team will have some nice sleep now. See you next week for some more ships.
OriginalAnd the team will have some nice sleep now. See you next week for some more ships.
Chinese translation团队现在终于可以好好睡一觉了。下周见，还会有更多发布。
Model context reading · Belgian Francophone phrasing lens“Ships” means product releases. This is a wrap-up note about the team resting and more releases next week, not a quota or reset commitment.
PreviewBecause we are beyond happy to have Astra rolled out today ahead of schedule and you have been super patient with us (not really, but it’s ok!)… we will do the full banked reset today too for all Plus, Pro and Business…
OriginalBecause we are beyond happy to have Astra rolled out today ahead of schedule and you have been super patient with us (not really, but it’s ok!)… we will do the full banked reset today too for all Plus, Pro and Business users. Lands end of day.
Happy Astra day and enjoy a phenomenal weekend.
PS: If you create the account or upgrade before 8pm PT you will get it too. Still time!
Chinese translation因为我们非常高兴 Astra 今天提前上线，而且你们对我们超级有耐心（其实也没有，但没关系！）……我们今天也会为所有 Plus、Pro 和 Business 用户做完整的 banked reset。今天结束前到账。 Astra 日快乐，祝你们周末非常棒。 附：如果你在太平洋时间晚上 8 点前创建账户或升级，也会拿到。还有时间！
Phrasesfull banked reset · Lands end of day · you will get it too
Model context reading · Belgian Francophone phrasing lens“Banked reset” reads as an account entitlement or reset credit that can land in users’ accounts, not a forced live usage/rate-limit reset. “Lands end of day” gives the grant timing, while “Happy Astra day” is launch celebration wording.
PreviewOK nevermind, the team and Astra did a good job and our systems are more scalable than we anticipated. Astra is now rolled out to all Plus and Business users too. Hope you have a blast and let us know how it goes!
OriginalOK nevermind, the team and Astra did a good job and our systems are more scalable than we anticipated.
Astra is now rolled out to all Plus and Business users too. Hope you have a blast and let us know how it goes!
Chinese translation好吧当我没说，团队和 Astra 做得很好，我们的系统比预期更可扩展。 Astra 现在也已经向所有 Plus 和 Business 用户推出了。希望你们玩得开心，也请告诉我们使用感受！
Phrasessystems are more scalable · rolled out to all Plus and Business users too
Model context reading · Belgian Francophone phrasing lensThis is a rollout and capacity update. “systems are more scalable” refers to infrastructure handling demand, and “rolled out” means Astra access expanded to Plus and Business users; it does not promise a Codex usage or rate-limit reset.
PreviewAwesome to see GPT-6 Astra is #1 on Terminal Bench 4.0 using the Codex harness. At 50% of the cost of #2. https://t.co/QeY4FfEcxc
OriginalAwesome to see GPT-6 Astra is #1 on Terminal Bench 4.0 using the Codex harness. At 50% of the cost of #2. https://t.co/QeY4FfEcxc
Chinese translation很高兴看到 GPT-6 Astra 使用 Codex harness 在 Terminal Bench 4.0 上排名第一。而且成本只有第二名的 50%。https://t.co/QeY4FfEcxc
Phrasesusing the Codex harness · At 50% of the cost of #2
Model context reading · Belgian Francophone phrasing lensThis is a benchmark and cost-efficiency post. The “cost” wording is about model or benchmark operating cost, not user quota, usage cycles, or reset credits.
PreviewSome Plus and Business users won't yet get access to Astra today, we've got you covered with a banked reset. Lands by end of day and if you create your account by 8pm PT then you'll get it too.
OriginalSome Plus and Business users won't yet get access to Astra today, we've got you covered with a banked reset. Lands by end of day and if you create your account by 8pm PT then you'll get it too.
Chinese translation一些 Plus 和 Business 用户今天还不能获得 Astra 访问权限，我们会用一个 banked reset 给你们补上。今天结束前到账；如果你在太平洋时间晚上 8 点前创建账号，也会拿到它。
Phrasesbanked reset · Lands by end of day · you'll get it too
Model context reading · Belgian Francophone phrasing lens“Banked reset” reads as a saved reset credit/entitlement for users missing Astra access today. It is not a live global Codex usage or rate-limit reset window.
Preview@BradGroux Congratulations. What will you do with it?
Original@BradGroux Congratulations. What will you do with it?
Chinese translation@BradGroux 恭喜。你打算用它做什么？
Model context reading · Belgian Francophone phrasing lensA normal reply to a user who received Astra access. In context, “it” refers to Astra access, not a reset, usage limit, or rate-limit event.
