/limits — Usage limits by plan — all providers
$ Every plan we track, side by side. Prices and allowances come from official docs and public reporting; the "last rule change" column links to the record. Click a provider for reset history.
Codex & ChatGPT Work · tokens
| Plan | Price | 5h gate | Allowance |
|---|---|---|---|
| Free / Go | $0 | Active | Enough to try it out |
| Plus | $20/mo | Active | Entry allowance; heavy debugging or large file reads hit the cap quickly |
| Pro | $100/mo | Disabled | About 5× Plus |
| Pro | $200/mo | Disabled | About 20× Plus, sized for a full working day |
| Business / Enterprise | per seat | Active | Each seat defaults to roughly Plus level |
Last rule change: 2026-09-10 · OpenAI pauses new subscriptions to the $200 Pro plan. The stated reason is that this tier puts the most strain…
Since 2026-04 usage is metered in tokens rather than message count, so no fixed message cap is published.
Claude Code · tokens
| Plan | Price | 5h gate | Allowance |
|---|---|---|---|
| Free | $0 | Active | Enough to try it out; no Claude Code |
| Pro | $20/mo | Active | At least 5× the free tier per session |
| Max 5x | $100/mo | Active | About 5× Pro per session; Opus has its own weekly cap |
| Max 20x | $200/mo | Active | About 20× Pro per session; Opus has its own weekly cap |
| Team / Enterprise | per seat | Active | Allocated per seat; also got the 5-hour doubling from 2026-05 |
Last rule change: 2026-09-01 · Fable 5.1 keeps the same pricing as Fable 5, but cache-read pricing drops from $1/MTok to $0.25/MTok. For long…
Only relative multiples are published, never absolute token counts; usage is metered on input + output tokens combined.
Grok Build · compute
| Plan | Price | 5h gate | Allowance |
|---|---|---|---|
| Free | $0 | — | Chat and Voice have their own allowance, outside the paid weekly pool |
| SuperGrok Lite | ~$10/mo | — | Entry-level weekly pool |
| SuperGrok | ~$30/mo | — | Full weekly pool, shared across Chat / Imagine / Voice / Build |
| SuperGrok Plus | ~$100/mo | — | Between the $30 and $300 tiers, noticeably larger weekly pool |
| SuperGrok Heavy | ~$300/mo | — | Largest consumer weekly pool |
Last rule change: 2026-06-01 · xAI drops the per-message daily cap in favour of a single compute-based weekly pool shared across Chat, Imagin…
No per-tier allowance is published — Settings → Usage only shows a percentage. Prices are from public reporting.
Gemini CLI · requests
| Plan | Price | 5h gate | Allowance |
|---|---|---|---|
| Free (personal Google account) | $0 | Active | 60 model requests per minute, 1,000 per day |
| Google AI Pro | ~$19.99/mo | Active | 1,500 model requests per user per day |
| Google AI Ultra | from ~$99.99/mo | Active | 2,000 model requests per user per day, plus more AI credits |
Last rule change: 2026-06-18 · For free-tier and Google One users, the Gemini CLI is replaced by the Antigravity CLI. Paid Google AI Pro / Ul…
The quota counts model requests — in agent mode a single prompt can trigger dozens internally. Prices are from public reporting.
Kimi Code · tokens
| Plan | Price | 5h gate | Allowance |
|---|---|---|---|
| Adagio | ¥0/mo | Active | K2.7 / K2, 2 projects |
| Andante | ¥49/mo | Active | About 300–1,200 calls per 5 hours, refreshed weekly |
| Moderato | ¥99/mo | Active | K3 flagship + K3-256K, 2 concurrent agents |
| Allegretto | ¥199/mo | Active | K3, 40 agent-cluster runs per month |
| Allegro | ¥699/mo | Active | K3 million-token conversations, 4 concurrent agents |
Last rule change: 2026-07-16 · Moonshot AI releases Kimi K3 (2.8 trillion parameters, up to a 1M-token context), rolled out the same day to K…
Allowance refreshes every 7 days from your subscription date and does not roll over; all signed-in devices and API keys share one pool.
GLM Coding Plan · prompts
| Plan | Price | 5h gate | Allowance |
|---|---|---|---|
| Lite (new plan) | ¥49/mo | Active | About 80 prompts per 5 hours, ~400 per week |
| Pro (new plan) | ¥149/mo | Active | About 400 prompts per 5 hours, ~2,000 per week |
| Max (new plan) | ¥469/mo | Active | About 1,600 prompts per 5 hours, ~8,000 per week |
| Legacy plan (before 2026-02-12) | ¥40 / 200 / 400 | Active | 1.5× the 5-hour allowance and no weekly cap |
Last rule change: 2026-02-12 · The GLM Coding Plan moves to new tiers and introduces a weekly cap for the first time. Existing subscribers fr…
Metered in prompts — not requests, not tokens. Monthly totals are extrapolated from the weekly cap and are not official figures.
Cursor · USD
| Plan | Price | 5h gate | Allowance |
|---|---|---|---|
| Hobby | $0 | — | No third-party model allowance |
| Pro | $20/mo | — | Includes $20 of third-party model usage at API rates |
| Pro+ | $60/mo | — | Includes $70 of third-party model usage |
| Ultra | $200/mo | — | Includes $400 of third-party model usage |
| Teams standard seat | $40/seat/mo | — | First-party models and third-party API in two separate pools |
| Teams Premium seat | $120/seat/mo | — | Added 2026-06, higher usage ceiling |
Last rule change: 2026-07-21 · Cursor doubles usage limits outright on every individual and team plan. The new limits apply to Grok, Composer…
You get a dollar allowance converted at API rates, not a request count. First-party and third-party models draw from separate pools.
Providers change these without notice. If a row is stale, the source link on each provider page has the current official figure — corrections welcome.
// NOTIFY ME
Email me when any provider resets
One email per official reset. No digest, no marketing, one-click unsubscribe. Prefer X? Reset alerts are also posted at @XIAOMIAO_26 · RSS