Claude limit reach? Wetin you fit do next

Which Claude limit you hit

A Claude limit reached message dey tell you which window you dey wait for, and na that window go determine wetin you fit do next. Session and weekly limits cover all models at once, so switching models no go restore your access. If na limit wey name one model, na exception: change model and continue your work while that one model remain blocked.

Claude Code dey print the window and its reset time for the same line:

You've hit your session limit · resets 3:45pm
You've hit your weekly limit · resets Mon 12:00am
You've hit your Opus limit · resets 3:45pm

The session window dey roll across five hours. Weekly windows dey reset for one fixed time every week wey dem assign to your account, so the day and hour remain the same from one week to the next. Pro plan get the five-hour session window plus one weekly limit wey apply across all models. Max plan add a second weekly limit wey apply to one model family alone. The actual ceilings dey inside the complete breakdown of Claude usage limits. This page dey start from where that one stop.

Na this part dey catch many people. Every request dey count against the session window and weekly window at the same time, so one heavy afternoon fit finish the weekly allowance while the five-hour window still get space. Check the reset time before you plan the rest of your day. Reset at 3:45pm fit be small break. Reset on Monday at midnight no be so.

Check the meter before you make guess

For claude.ai, open Settings, then Usage. Progress bars go show how much of the five hour session window and each weekly window you don use. So you go know which one don finish instead of guessing.

For Claude Code, run /usage. E go show the same plan bars, plus breakdown of where recent usage go. E go attribute am to skills, subagents, plugins, and individual MCP (model context protocol) servers as percentage of the total. Press d for the last 24 hours and w for the last 7 days. E also go flag any pattern wey account for 10% or more of recent usage. The two patterns you go see pass na long context and cache misses.

Two things you need note. The breakdown dey calculate from session history wey dey stored for that machine. So work wey you do for another laptop or for claude.ai no dey inside. Also, the dollar figure for the Session block for top dey use standard list rates for API (application programming interface) users. This means say e no be your bill when you dey use subscription.

Smaller model dey help?

E only help when message mention a model. After You've hit your Opus limit, run /model, choose Sonnet or Haiku, and the session go continue with everything wey dey context. After You've hit your session limit or You've hit your weekly limit, /model no change anything, because all models dey share those windows.

The switch dey worth much more before you reach the limit than after. Sonnet fit handle most coding work for small part of Opus rate, so to leave Opus as your default na one of the two habits behind most unexpected usage. The other one na session wey you never clear. Give mechanical edits and log triage to Haiku, and keep Opus for the design decision wey really dey hold you back. Choosing between Opus, Sonnet and Haiku explain where each one justify the price.

Stop dey pay for context wey you don stop to use

Claude Code dey send your complete conversation with every request. Each time Claude use tool, e dey send another request wey carry that batch of tool results. Prompt caching mean say system go read the history again for cached rate instead of full input rate. This make am cheaper, but e no make am free. So one-line question for session wey you don leave open since morning still dey use resources for the complete conversation.

/context
/clear
/compact Focus on the failing tests and the files we changed

/clear no cost anything because e dey throw the context away instead of reading am. /compact must read the conversation wey e summarize, so compacting very large context na itself a large request. Use /clear when you move go unrelated work, and use /compact only when you dey inside task and need make the thread continue. Run /rename before you clear am, so you fit find the session again with /resume.

Cache lifetime matter pass wetin many people expect. For subscription, cache dey live for one hour. E drop to five minutes once you start use usage credits, and five minutes na also the default for API key. Your first message after long lunch no dey find cache, so e process the complete context again for full input price. Na why usage fit jump without any visible reason. Where prompt caching dey pay for itself explain the arithmetic.

Three more levers dey for the same place. /mcp list the servers wey you configure, so you fit switch off the ones wey this task no need. Long CLAUDE.md dey load when session start, whether the work need am or not. So move specialized instructions enter skills, wey dey load only when you invoke dem. Extended thinking dey bill as output tokens, and na the expensive half. So reduce am with /effort, off am for /config, or set MAX_THINKING_TOKENS=8000 for model wey get fixed thinking budget.

Divide the work so e fit enter one window

Press Shift+Tab to switch enter plan mode before you make big change. Claude go inspect the codebase and suggest one approach for you to approve. If the direction no correct, catching am for there go cost one plan instead of two hours of edits plus the rework.

Give noisy operations to subagents. If you run test suite or read 10,000-line log inside subagent, the raw output go remain for that subagent context, and na only the summary go return to your main conversation.

Press Escape immediately when Claude start dey go wrong direction. /rewind go restore both the conversation and the code to earlier checkpoint, so you no go pay to undo work wey you never want.

You suppose buy usage credits or use the API?

Usage credits let Pro or Max account continue to work after plan allowance don finish, and dem go bill am with standard API rates. Turn am on for Settings, then Usage for claude.ai after you attach payment method, or run /usage-credits for Claude Code. E go open your billing settings for browser. Credits apply to Claude conversations and Claude Code terminal usage, with daily redemption limit of $2000. The command need subscription login, so e no dey available when you authenticate with API key. For free plan, credits no dey available at all. So the first step na Pro at $20 per month. wetin Pro add and where its limits fit stop you explain how far that one fit carry you. If you dey add credits most weeks, bigger allowance usually cheaper. the difference between the $100 and $200 Max tiers na usage multiplier, no be difference for models or features.

The other option na plain API key from Claude Console. Dem bill am per token, with no session or weekly window. One problem fit happen: stray ANTHROPIC_API_KEY for your environment go route Claude Code through that key instead of your subscription, without telling you. You go discover am from the bill, or from this:

API Error: Request rejected (429) · this may be a temporary capacity issue. If it persists, check https://status.claude.com.

That 429 na the key rate limit, no be your plan limit. Run /status and confirm say the active credential na the one wey you intend to use.

Wetin one afternoon for API go cost?

The worked example below na one afternoon of agent-style coding: 200,000 uncached input tokens, 1.8 million cache reads, and 150,000 output tokens. Prices na Anthropic list rates as of August 2026. Every one of these rates na per million tokens, so if that unit still dey look abstract, wetin one million tokens really worth go turn am to amount of reading and writing wey you fit picture.

[
  {
    "label": "Opus 5",
    "input_usd_per_mtok": 5,
    "cache_read_usd_per_mtok": 0.5,
    "output_usd_per_mtok": 25,
    "afternoon_usd": 5.65
  },
  {
    "label": "Sonnet 5",
    "input_usd_per_mtok": 2,
    "cache_read_usd_per_mtok": 0.2,
    "output_usd_per_mtok": 10,
    "afternoon_usd": 2.26
  },
  {
    "label": "Haiku 4.5",
    "input_usd_per_mtok": 1,
    "cache_read_usd_per_mtok": 0.1,
    "output_usd_per_mtok": 5,
    "afternoon_usd": 1.13
  }
]

The same afternoon costs $5.65 for Opus 5 and $1.13 for Haiku 4.5. Output na wetin dey move the total: at $25 per million tokens for Opus 5 compared with $5 for Haiku 4.5, the 150,000 tokens wey Claude write cost pass the 2 million tokens wey e read. Na why reducing the thinking budget save pass trimming files.

Make you note one date. Sonnet 5 dey use introductory pricing of $2 and $10 per million input and output tokens until 31 August 2026, then e go change to $3 and $15 on 1 September 2026. This go push the same afternoon from $2.26 to about $3.39.

For scale, Anthropic own documentation talk say average Claude Code spend across enterprise deployments dey near $13 per developer per active day, and e dey below $30 per active day for 90% of users, as of August 2026. Per-token billing no automatically cheaper than subscription. E dey win when your load dey come in spikes, because quiet days cost nothing, and e dey lose when you work every day. API compared with subscription, using real numbers get the crossover point.

Run the long job for one VPS overnight

Some jobs no need make you dey watch dem: big refactor, test migration, documentation sweep. Those jobs no need your subscription window too. Put dem for one VPS (virtual private server) with an API key. Dem go use per token billing while your plan allowance remain untouched for tomorrow interactive work.

Install Claude Code for one Ubuntu box and confirm say e dey run:

curl -fsSL https://claude.ai/install.sh | bash
claude --version

claude --version suppose print version string like 2.1.211 (Claude Code). If you see command not found here, e mean say ~/.local/bin never dey your PATH, so open new shell and try again.

Put the key for file instead of your shell history. Create ~/.claude-env with editor and put one line inside, export ANTHROPIC_API_KEY=sk-ant-your-key-here. Then lock the file down and load am:

chmod 600 ~/.claude-env
. ~/.claude-env
printenv ANTHROPIC_API_KEY | cut -c1-10

That last line dey print the first ten characters. This one enough to confirm say the variable dey set without showing the complete key for screen.

Now start the job inside tmux so e go continue even if your laptop close:

tmux new -s overnight
claude --bare -p "Convert every test under ./tests from unittest to pytest, run the suite, and fix what fails" \
  --allowedTools "Read,Edit,Bash" \
  --append-system-prompt-file ./CLAUDE.md \
  --output-format json > ~/overnight.json

Detach with Ctrl-b then d, and close the connection. Reattach later with tmux attach -t overnight. Claude Code for tmux session on VPS explain the session handling completely.

--bare important here. E skip auto-discovery of hooks, skills, plugins and CLAUDE.md. E no ever read OAuth credentials too, so the run go use the API key and nothing else. Na this one keep the job away from your subscription. E also mean say the project rules don disappear, na why --append-system-prompt-file dey return dem deliberately. Without --bare, any set ANTHROPIC_API_KEY go trigger one-time approval prompt wey unattended run no fit answer.

For morning, read the result and cost from the JSON:

jq -r '.result' ~/overnight.json
jq -r '.total_cost_usd' ~/overnight.json

Claude Code dey exit with 0 when e succeed, and non-zero when the run fail. So wrapper script fit branch based on the exit status and send the failure to you by mail. Set workspace spend limit for Claude Console before you leave am running. Agent wey dey retry loop for 3am fit spend real money. Reduce CLAUDE_CODE_MAX_TOOL_USE_CONCURRENCY if you also want slow am down. Cost control for agents wey you leave running go through the limits wey worth setting first.

Give the job its own unprivileged user and its own git branch. Agent wey dey edit repository unattended fit sometimes do something wey you for stop, and branch go make am a git checkout instead of incident. How to run coding agent safely for VPS cover the isolation.

For bulk work wey no interactive at all, like classifying thousands of records, Batch API dey process requests asynchronously with 50% discount for both input and output tokens. Na the cheapest per token option Anthropic dey publish, and overnight window na exactly the latency wey e require.

Wetin no go restore your access

Starting new conversation no go restore your allowance. Usage dey count against your account, no be against one conversation. E go make every message after that cheaper, because fresh conversation no get history wey e need read again, so e still make sense to do am. Na cost cut, no be reset.

Deleting old conversations no go refund anything. Dem tokens don spend when system process dem.

Changing model after session or weekly message no go change anything, because all models dey share those time windows. Na only message wey name model dey respond to /model.

Retrying am for loop no go do anything apart from keep your terminal busy. The reset time wey message show na the real answer, and /usage go show you the same clock without the error.

FAQ

How long until my Claude limit resets?

The message dey tell you. Session limits dey run for rolling five hour window, and the message go show clock time, like resets 3:45pm. Weekly limits dey reset for fixed time every week wey dem assign to your account, and the message go show the day, like resets Mon 12:00am. To see both bars and both reset times before you hit any limit, run /usage for Claude Code, or open Settings, then Usage for claude.ai.

Does switching to Sonnet get me working again?

Only when the message mention one model. After You've hit your Opus limit, run /model, choose another model, and the session go continue with your context intact. After You've hit your session limit or You've hit your weekly limit, model choice no go make difference, because those windows dey count every model together. If you switch to cheaper model before you hit the limit, na that one really dey make the window last longer.

Will starting a new chat reset my usage limit?

No. Usage dey count against your account, so new chat go start with the same empty allowance. But e go make each next message cheaper, because new conversation no carry history wey Claude need read again. /clear for Claude Code dey do the same thing and e cost nothing, while /compact need read the conversation wey e summarize, so compacting very large session na expensive request by itself.

Is the API cheaper than upgrading my plan?

E depend on how steady your usage dey. The afternoon work wey this page calculate costs $2.26 for Sonnet 5 and $1.13 for Haiku 4.5 at August 2026 list rates. Subscription na flat fee for capped allowance, so e dey better when you work most days. Per token billing dey better when your load dey spike, because quiet days no go cost you anything.

Can I keep a job running while I wait for the reset?

Yes, if you move am away from your subscription. Run am for VPS with ANTHROPIC_API_KEY set for the environment, and billing go dey per token instead of drawing from your plan window, so both no go compete. claude --bare -p inside tmux go survive when your laptop close, and --output-format json go write one total_cost_usd field wey you fit check for morning with jq.