LimitHit

Live from vendor pages · honest unknowns

Hit your limit? So did everyone.

Every AI coding plan's price, usage limits and reset window on one page, with a changelog for when they quietly move.

Where vendors don't publish a number, we say unknown instead of guessing.

claude --continue

Error: usage limit reached. Try again later.

Heads up ·

Codex Pro 200 grandfathered allowance ends Oct 29

Subscribers active Sep 22–29 keep the old, bigger allowance only through Oct 29 (OpenAI confirms the date). The reported size of the drop (20× → 10× Plus) comes from a third-party report, not OpenAI.

Codex limits →

Top 5 bang for buck

How much API usage each $1 of subscription buys. Claude Pro: $20/mo gets you $738–1,173 of API usage.

Used to the limit every week · ranked by the low estimate, then the high

  • low estimate
  • up to the high estimate
  • how sure we are (high, medium, low)

API usage per $1 of subscription (× price)

  1. API usage worth 37–59× the price,

    Claude

    Pro

    $20/mo

    medium
  2. API usage worth 36–59× the price,

    Claude

    Max 5x

    $100/mo

    low
  3. API usage worth 36–59× the price,

    Claude

    Max 20x

    $200/mo

    medium
  4. API usage worth 10–14× the price,

    Codex

    Pro 100

    $100/mo

    low
  5. API usage worth 10–14× the price,

    Codex

    Pro 200

    $200/mo

    low
Show the numbers as a table
Top 5 plans by API usage per $1 of subscription: the data behind the chart, in chart order.
PlanAPI value /mo× price
Claude Pro$20/momedium$738–$1,17337×–59×
Claude Max 5x$100/molow$3,600–$5,86336×–59×
Claude Max 20x$200/momedium$7,121–$11,72636×–59×
Codex Pro 100$100/molow$1,042–$1,44910×–14×
Codex Pro 200$200/molow$2,084–$2,89710×–14×

Column height = what a fully used month would cost at API list prices ÷ the plan's price. Of 18 plans with a published API value. Codex Pro 500 ties at 10× on the low end but misses the cut on the high end (14×). 26 publish none and aren't ranked.

All 18 plans by multiple →

All plans

API value* = a fully used month at API list prices · How we estimate it · every row cites its source

44 of 44 plans
$20/mo
Pro
$738–1,173/mo~37–59×mediumhow?
At least 5x Free usage per 5-hour session; weekly limits on top; no fixed message count
Rolling 5-hour session + weekly
$17/mo billed annually ($200/yr). Shared pool with Claude chat.
$100/mo
Max 5x
$3.6k–5.9k/mo~36–59×lowhow?
5x Pro usage per 5-hour session; weekly limits; Fable up to 50% of weekly limits; $100/mo API credits
Rolling 5-hour session + weekly
Monthly billing only. Exact token/message caps unknown.
$200/mo
Max 20x
$7.1k–11.7k/mo~36–59×mediumhow?
20x Pro usage per 5-hour session; weekly limits; Fable up to 50% of weekly limits; $200/mo API credits
Rolling 5-hour session + weekly
Price $200 per pricing table ('From $100'; $200 API credits … more
Price $200 per pricing table ('From $100'; $200 API credits tier). Exact caps unknown.
$20/mo
Plus
$162–253/mo~8.1–13×lowhow?
Est. local messages/5h: GPT-6.1 Sol 15-160, GPT-6 Sol 15-150, GPT-6 Astra 5-45 (limited), GPT-6 Luna 350-3,000; weekly limits may apply
5-hour + weekly
Estimates, not fixed limits.
Source for Codex Plus ↗
checked
$100/mo
Pro 100
$1k–1.4k/mo~10–14×lowhow?
No 5-hour limit; included allowance (weekly); relative size vs Plus not stated on official page
Weekly
third-partyThird parties report ~5x Plus (unverified officially).
$200/mo
Pro 200
$2.1k–2.9k/mo~10–14×lowhow?
No 5-hour limit; more than Pro 100; new subs get lower allowance than legacy Pro 200
Weekly
third-partyGrandfathered (active Sep 22-29 2026) keep old allowance thr… more
Grandfathered (active Sep 22-29 2026) keep old allowance through Oct 29 2026. ~10x vs 20x Plus is third-party/X-reported.
$500/mo
Pro 500
$5.2k–7k/mo~10–14×lowhow?
No 5-hour limit; highest included usage; Ultrafast (8x included usage rate, 6x credit rate)
Weekly
—
$25/mo
Business (Standard seat)
$162–253/mo~6.5–10×lowhow?
Same 5-hour estimates as Plus; weekly may apply
5-hour + weekly
$25/user/mo billed monthly; $20/user/mo billed annually (2+ … more
$25/user/mo billed monthly; $20/user/mo billed annually (2+ users), per developers.openai.com/codex/pricing. A $100 Business tier is mentioned on chatgpt.com/codex/pricing, details unknown.
Free
Free
not published
Limited agent/chat requests
 unknown
—
$8/mo
Plus
not published
Basic pool shared across Cursor and Grok
Weekly refresh
—
$30/mo
Premium
not published
4x Plus usage; third-party models at API list price, not included
Weekly refresh
—
$100/mo
Super
not published
15x Plus; priority at peak
Weekly refresh
—
$200/mo
Ultra
not published
50x Plus; earliest access
Weekly refresh
—
$20/mo
Pro (legacy $20)
not published
 unknown
 unknown
third-partyNot on official pricing page anymore; reportedly closed to n… more
Not on official pricing page anymore; reportedly closed to new signups Oct 12 2026 (third-party).
$60/mo
Pro+ (legacy)
not published
 unknown
 unknown
Legacy plan: keeps working for existing subscribers; new cus… more
Legacy plan: keeps working for existing subscribers; new customers start on The X Subscription. Cursor lists Super ($100/mo) as the comparable tier. Includes Cursor Models + Other Models pools (more than Pro); pools reset with the monthly billing cycle.
Free
Free
not published
Light agent quota; unlimited Tab completions and inline edits
Daily + weekly (calendar date)
No extra usage: when the quota runs out, wait for the next d… more
No extra usage: when the quota runs out, wait for the next daily or weekly reset. Free models don't count against quota.
Source for Devin Free ↗
checked
$20/mo
Pro
not published
Daily + weekly quota shared across Devin Desktop, Devin Cloud and Devin CLI; size not published
Daily + weekly (calendar date)
Quota is token-based: cost per token varies by model. The da… more
Quota is token-based: cost per token varies by model. The daily quota is more than 1/7 of the weekly. Extra usage at API list prices from prepaid on-demand credits (never expire). Up to 10 concurrent cloud sessions. Windsurf's Mar 2026 estimate: 7–27 Premium Plus (Opus 4.6, GPT-5.4) or 47–190 lightweight messages/day.
Source for Devin Pro ↗
checked
$200/mo
Max
not published
'Significantly higher' weekly quota than Pro, no daily cap; size not published
Weekly (calendar date)
Daily cap removed 2026-04-06. Extra usage at API list prices… more
Daily cap removed 2026-04-06. Extra usage at API list prices. Up to 10 concurrent cloud sessions. Windsurf's Mar 2026 estimate (when Max still had a daily window): 42–170 Premium Plus messages/day.
Source for Devin Max ↗
checked
$40/mo
Teams (full seat)
not published
Pro-equivalent daily + weekly quota per full seat; on-demand credits shared by the team
Daily + weekly (calendar date)
$80/month team minimum (two full seats, or the rest paid as … more
$80/month team minimum (two full seats, or the rest paid as prepaid credits). Up to 200 users. Free flex seats draw only from shared credits and don't include Devin Desktop. Unlimited concurrent sessions.
$15/mo
Windsurf Pro (legacy $15)Closed to new signups
not published
Same quota system as Pro
Daily + weekly (calendar date)
Existing Windsurf Pro subscribers keep $15/mo indefinitely s… more
Existing Windsurf Pro subscribers keep $15/mo indefinitely since the 2026-03-19 switch to quotas. Not sold to new users.
$30/mo
Windsurf Teams (legacy $30/seat)Closed to new signups
not published
Same quota system as Teams
Daily + weekly (calendar date)
Existing Windsurf Teams subscribers keep $30 per Developer s… more
Existing Windsurf Teams subscribers keep $30 per Developer seat indefinitely. Not sold to new users.
Free
Free
not published
Allowance of AI credits (amount not stated); 2,000 completions/month
Monthly
—
Source for Copilot Free ↗
checked
$10/mo
Pro
$15/mo~1.5×highhow?
1,000 base + 500 flex = 1,500 AI credits/mo (1 credit=$0.01); unlimited completions
Monthly
—
Source for Copilot Pro ↗
checked
$39/mo
Pro+
$70/mo~1.8×highhow?
3,900 base + 3,100 flex = 7,000 AI credits/mo; unlimited completions
Monthly
—
Source for Copilot Pro+ ↗
checked
$100/mo
Max
$200/mo~2×highhow?
10,000 base + 10,000 flex = 20,000 AI credits/mo
Monthly
—
Source for Copilot Max ↗
checked
$19/mo
Business
$19/mo~1×highhow?
1,900 AI credits/user/mo pooled; overage $0.01/credit
Monthly
—
$22.80/mo
StandardClosed to new signups
not published
1,500 agent-mode/CLI requests/user/day; 6,000 code requests/day; 960 chat requests/day; 2 req/s
Daily (Google Cloud quota reset)
$22.80/user/mo monthly; $19/user/mo with upfront annual comm… more
$22.80/user/mo monthly; $19/user/mo with upfront annual commitment (codeassist.google). New subscriptions no longer sold since 2026-10-09; existing ones auto-renew through the rest of 2026, auto-renewal ends in 2027 (GCA release notes). Quotas from developers.google.com/gemini-code-assist/resources/quotas.
$54/mo
EnterpriseClosed to new signups
not published
2,000 agent-mode/CLI requests/user/day
Daily
$54/user/mo monthly; $45/user/mo with upfront annual commitm… more
$54/user/mo monthly; $45/user/mo with upfront annual commitment (codeassist.google). New subscriptions no longer sold since 2026-10-09; existing ones auto-renew through the rest of 2026, auto-renewal ends in 2027 (GCA release notes). Quotas from developers.google.com/gemini-code-assist/resources/quotas.
Free
Free / individual (Google login)Retired
not published
 unknown
 unknown
Retired 2026-06-18: for the unpaid tier (Gemini Code Assist … more
Retired 2026-06-18: for the unpaid tier (Gemini Code Assist for individuals) and Google One users, Gemini CLI and the Code Assist IDE extensions were replaced by Antigravity CLI / Antigravity (banners on codeassist.google and geminicli.com).
$18/mo
Lite
$71–151/mo~3.9–8.4×mediumhow?
10,000 credits per week + 2,000 per 5 hours. Credits = tokens x model multiplier (GLM-5.3 per 10k tokens: input 6.9, cached 1.7, output 24 credits). Off-peak usage costs 50% of the credits; peak is Mon-Fri 14:00-18:00 UTC+8 (23:00-03:00 PT), weekends are off-peak all day.
Rolling 5-hour + weekly (7 days from subscription)
$14.4/mo billed quarterly, $12.6/mo billed yearly. Quarterly… more
$14.4/mo billed quarterly, $12.6/mo billed yearly. Quarterly billing 20% off, yearly 30% off. Credits system since 2026-07-30; legacy V1/V2 prompt-based plans still renew for existing subscribers. Non-refundable. Works in Claude Code, Codex, OpenCode, Cline, Kilo Code and 20+ other agent tools (Anthropic-compatible endpoint).
$80/mo
Pro
$425–907/mo~5.3–11×mediumhow?
60,000 credits per week + 12,000 per 5 hours (6x Lite). Credits = tokens x model multiplier (GLM-5.3 per 10k tokens: input 6.9, cached 1.7, output 24 credits). Off-peak usage costs 50% of the credits; peak is Mon-Fri 14:00-18:00 UTC+8 (23:00-03:00 PT), weekends are off-peak all day.
Rolling 5-hour + weekly (7 days from subscription)
$64/mo billed quarterly, $56/mo billed yearly. Quarterly bil… more
$64/mo billed quarterly, $56/mo billed yearly. Quarterly billing 20% off, yearly 30% off. Credits system since 2026-07-30; legacy V1/V2 prompt-based plans still renew for existing subscribers. Non-refundable. Works in Claude Code, Codex, OpenCode, Cline, Kilo Code and 20+ other agent tools (Anthropic-compatible endpoint).
$168/mo
Max
$992–2,117/mo~5.9–13×mediumhow?
140,000 credits per week + 28,000 per 5 hours (14x Lite). Credits = tokens x model multiplier (GLM-5.3 per 10k tokens: input 6.9, cached 1.7, output 24 credits). Off-peak usage costs 50% of the credits; peak is Mon-Fri 14:00-18:00 UTC+8 (23:00-03:00 PT), weekends are off-peak all day.
Rolling 5-hour + weekly (7 days from subscription)
$134.4/mo billed quarterly, $117.6/mo billed yearly. Quarter… more
$134.4/mo billed quarterly, $117.6/mo billed yearly. Quarterly billing 20% off, yearly 30% off. Credits system since 2026-07-30; legacy V1/V2 prompt-based plans still renew for existing subscribers. Non-refundable. Works in Claude Code, Codex, OpenCode, Cline, Kilo Code and 20+ other agent tools (Anthropic-compatible endpoint).
$19/mo
Plus
not published
Rolling 5-hour usage limit (no weekly limit on new plans) plus a monthly credit pool shared with all Kimi features (Agent, Kimi Work, Claw). Token/request caps not published. K3.1 Pro and K3 included; no 1M context on them, no Highspeed.
Rolling 5-hour + monthly credit pool refresh
$180/yr billed annually. China SKU ¥99/mo (separate price, n… more
$180/yr billed annually. China SKU ¥99/mo (separate price, not a conversion). Go tier (China only) excludes Kimi Code. Legacy plans (Andante to Allegro) keep 5-hour + weekly limits. All devices and API keys share one quota; works in Kimi Code CLI, VS Code, Claude Code, OpenCode.
$39/mo
Pro
not published
Rolling 5-hour usage limit (no weekly limit on new plans) plus a monthly credit pool shared with all Kimi features (Agent, Kimi Work, Claw). Token/request caps not published. K3.1 Pro and K3 with 1M context + Highspeed. Highspeed uses 2× quota.
Rolling 5-hour + monthly credit pool refresh
$372/yr billed annually. China SKU ¥199/mo (separate price, … more
$372/yr billed annually. China SKU ¥199/mo (separate price, not a conversion). Go tier (China only) excludes Kimi Code. Legacy plans (Andante to Allegro) keep 5-hour + weekly limits. All devices and API keys share one quota; works in Kimi Code CLI, VS Code, Claude Code, OpenCode.
$99/mo
Max
not published
Rolling 5-hour usage limit (no weekly limit on new plans) plus a monthly credit pool shared with all Kimi features (Agent, Kimi Work, Claw). Token/request caps not published. K3.1 Pro and K3 with 1M context + Highspeed. Highspeed uses 2× quota. Larger credit pool than Pro (size not published).
Rolling 5-hour + monthly credit pool refresh
$948/yr billed annually. China SKU ¥699/mo (separate price, … more
$948/yr billed annually. China SKU ¥699/mo (separate price, not a conversion). Go tier (China only) excludes Kimi Code. Legacy plans (Andante to Allegro) keep 5-hour + weekly limits. All devices and API keys share one quota; works in Kimi Code CLI, VS Code, Claude Code, OpenCode.
$199/mo
Ultra
not published
Rolling 5-hour usage limit (no weekly limit on new plans) plus a monthly credit pool shared with all Kimi features (Agent, Kimi Work, Claw). Token/request caps not published. K3.1 Pro and K3 with 1M context + Highspeed. Highspeed uses 2× quota. Largest credit pool (size not published).
Rolling 5-hour + monthly credit pool refresh
$1,908/yr billed annually. China SKU ¥1,399/mo (separate pri… more
$1,908/yr billed annually. China SKU ¥1,399/mo (separate price, not a conversion). Go tier (China only) excludes Kimi Code. Legacy plans (Andante to Allegro) keep 5-hour + weekly limits. All devices and API keys share one quota; works in Kimi Code CLI, VS Code, Claude Code, OpenCode.
$22/mo
Go
$131–178/mo~6–8.1×lowhow?
1x (baseline). No video model. 5-hour window + weekly window; text, image, audio and video share one quota. Absolute size not published.
Fixed 5-hour window + 7-day window, each starting at first use
Annual = 10 months' price ($220/yr). M Plan replaced Token P… more
Annual = 10 months' price ($220/yr). M Plan replaced Token Plan (Plus/Max/Ultra) for new buyers; MiniMax says text limits are unchanged. Works in MiniMax Code, Claude Code, Codex, Cursor, OpenCode via a Subscription Key. Credit packs cover overflow.
$55/mo
Explore
$393–534/mo~7.1–9.7×lowhow?
3x Go usage. H3 video included (weekly window only). 5-hour window + weekly window; text, image, audio and video share one quota. Absolute size not published.
Fixed 5-hour window + 7-day window, each starting at first use
Annual = 10 months' price ($550/yr). M Plan replaced Token P… more
Annual = 10 months' price ($550/yr). M Plan replaced Token Plan (Plus/Max/Ultra) for new buyers; MiniMax says text limits are unchanged. Works in MiniMax Code, Claude Code, Codex, Cursor, OpenCode via a Subscription Key. Credit packs cover overflow.
$132/mo
Build
$963–1,309/mo~7.3–9.9×lowhow?
7.5x Go usage. H3 video included (weekly window only). 5-hour window + weekly window; text, image, audio and video share one quota. Absolute size not published.
Fixed 5-hour window + 7-day window, each starting at first use
Annual = 10 months' price ($1,320/yr). M Plan replaced Token… more
Annual = 10 months' price ($1,320/yr). M Plan replaced Token Plan (Plus/Max/Ultra) for new buyers; MiniMax says text limits are unchanged. Works in MiniMax Code, Claude Code, Codex, Cursor, OpenCode via a Subscription Key. Credit packs cover overflow.
$50/mo
ProClosed to new signups
not published
6,000 requests per 5 hours, 45,000 per week, 90,000 per month (a request = one model call)
Rolling 5-hour + weekly (Mon 00:00 UTC+8) + monthly
Closed to new buyers: Alibaba says the limited-quantity Codi… more
Closed to new buyers: Alibaba says the limited-quantity Coding Plan Pro 'is no longer available once sold out' and recommends Token Plan, and the AI Coding Plan campaign page now redirects to the Token Plan page (checked 2026-10-20). Coding Plan Lite stopped new purchases 2026-03-20 and renewals 2026-04-13.
$8/mo
Lite
not published
11,500 credits per subscription month; 1-2 concurrent agents. Credits per token not published. Extra bundles $15 = 20,000 credits (max 5).
Monthly (subscription month)
List price shown; limited-time sale $6/mo, $65/yr (list $96)… more
List price shown; limited-time sale $6/mo, $65/yr (list $96). Temporary 2x usage per credit on Qwen-3.8-Max and DeepSeek-V4-Pro-0813. Singapore region only. General AI subscription that works in Claude Code, Qwen Code, Cursor, Codex, Qoder. Team seats: Standard $30, Pro $100, Max $200 per seat/mo (list).
$16/mo
Essential
not published
25,500 credits per subscription month; 2-3 concurrent agents. Credits per token not published. Extra bundles $15 = 20,000 credits (max 5).
Monthly (subscription month)
List price shown; limited-time sale $10/mo, $100/yr (list $1… more
List price shown; limited-time sale $10/mo, $100/yr (list $192). Temporary 2x usage per credit on Qwen-3.8-Max and DeepSeek-V4-Pro-0813. Singapore region only. General AI subscription that works in Claude Code, Qwen Code, Cursor, Codex, Qoder. Team seats: Standard $30, Pro $100, Max $200 per seat/mo (list).
$25/mo
Standard
not published
45,000 credits per subscription month; 3-4 concurrent agents. Credits per token not published. Extra bundles $15 = 20,000 credits (max 5).
Monthly (subscription month)
List price shown; limited-time sale $18/mo, $195/yr (list $3… more
List price shown; limited-time sale $18/mo, $195/yr (list $300). Temporary 2x usage per credit on Qwen-3.8-Max and DeepSeek-V4-Pro-0813. Singapore region only. General AI subscription that works in Claude Code, Qwen Code, Cursor, Codex, Qoder. Team seats: Standard $30, Pro $100, Max $200 per seat/mo (list).
$80/mo
Pro
not published
180,000 credits per subscription month; 6-8 concurrent agents. Credits per token not published. Extra bundles $15 = 20,000 credits (max 5).
Monthly (subscription month)
List price shown; limited-time sale $68/mo. Temporary 2x usa… more
List price shown; limited-time sale $68/mo. Temporary 2x usage per credit on Qwen-3.8-Max and DeepSeek-V4-Pro-0813. Singapore region only. General AI subscription that works in Claude Code, Qwen Code, Cursor, Codex, Qoder. Team seats: Standard $30, Pro $100, Max $200 per seat/mo (list).

Latest changes All changes →

  1. OpenAI

    upcoming

    Upcoming: grandfathered Pro 200 allowance ends (reported 20x -> 10x Plus)

    Multipliers from third-party/X post, not official page

    Source ↗

  2. Moonshot AI

    Kimi K3.1 Pro API launched: $2 in / $0.20 cached / $10 out (K3: $3 / $0.30 / $15); Highspeed $4 / $0.20 / $20; 1M context

    Source ↗

  3. Moonshot AI

    Kimi Code adds K3.1 Pro (Plus and up) and K3.1 Pro Highspeed (Pro and up, 2× quota)

    Source ↗

Get alerts when limits change

One short email per real change. Nothing else.

Tools: all

Your email and picks are held by Resend, our email provider. Never sold. Unsubscribe from any email.

New: AI app builder credits, 5 tools compared →Also: AI code review, 11 tools →Free: AI stack cost calculator (.xlsx) →Explainer: the Claude usage-limit lawsuits →