Hit your Claude Code usage limit? How the limits work and what you can do

What the limits are (October 2026)

Only officially published facts are listed. Weekly amounts and message counts are not published; your account's Settings > Usage is the source of truth.

Providers can change limits at any time. Anthropic also reserves the right to limit usage in other ways (weekly/monthly caps, model or feature limits). Facts checked 2026-10-10.

Official sources:claude.com/pricingWhat is the Max plan? (Claude Help Center)Use Claude Code with your Pro or Max planCodex pricing (OpenAI)Using Codex with your ChatGPT plan (OpenAI Help)

Why Claude Code hits limits fast

1. 1

   Every turn resends the whole conversation

   Files read, tool output and diffs are counted again with each message. The longer a session runs, the heavier each turn gets.

2. 2

   Opus uses more than Sonnet

   Anthropic suggests Sonnet for routine coding and Opus for hard problems, or /model opusplan (Opus plans, Sonnet executes).

3. 3

   Subagents, agent teams and scheduled tasks each draw usage

   Anthropic's docs note that agent teams use about 7x the tokens in plan mode.

4. 4

   Cache misses

   Come back after the cache expires and your first message reprocesses the full context.

5. 5

   Growing context

   Large pastes, many MCP servers and a long CLAUDE.md make every turn bigger.

Save allowance: /clear between tasks, /compact mid-task, and /usage to see what is eating your allowance.

Your four options

No option is always best. Pick by how much you use and how soon you need to keep working.

Pay-as-you-go is billed per token, and rate and concurrency limits still apply as described in our docs. You can also set daily, monthly or total spend caps on each API key. docs.aicodewith.ai

The math: what a Claude Code workday costs

Below are three workday intensities priced at Opus 5.5 official API rates. Most Claude Code tokens are cache reads, so the cache-read price matters a lot.

The same usage on aicodewith

At current public prices, Claude models on aicodewith start from 19% of the official price (price depends on the channel; the pricing page is authoritative).

Claude model prices (per 1M tokens)

See Claude model prices|/pricing

Your real token mix can differ a lot. Run /usage or /cost in Claude Code and plug in your own numbers.

Switch Claude Code to aicodewith in 1 minute

Claude Code reads the API endpoint and key from environment variables. The config below is the same one the dashboard gives you.

1. 1. Create an API key

   Create an API key in the aicodewith dashboard. You can set a daily or monthly spend cap on it.

2. 2. Edit ~/.claude/settings.json

   Add the env block below (on Windows: %USERPROFILE%\.claude\settings.json). ANTHROPIC_BASE_URL is the bare domain, with no /v1.

   ~/.claude/settings.json

   {
     "env": {
       "ANTHROPIC_AUTH_TOKEN": "YOUR_API_KEY",
       "ANTHROPIC_API_KEY": "",
       "ANTHROPIC_BASE_URL": "https://api.aicodewith.ai"
     }
   }

3. 3. Restart and verify

   Restart Claude Code and send a message. Pick Opus 5.5 or Sonnet 5.5 with /model, then confirm the request in your aicodewith usage records.

If an API key is set, Claude Code uses API billing instead of your subscription allowance.

Limit messages, decoded

Claude Code messages are quoted from Anthropic's errors documentation.

FAQ