[
    {
        "role": "user",
        "content": "apparently%2C%20i%20need%20to%20increase%20the%20token%20size%20limit%20beyond%2030000%20for%20my%20claude%20api%20usage.%20where%20do%20i%20go%20to%20configure%20this%3F"
    },
    {
        "role": "assistant",
        "content": "You currently can\u2019t increase the token-size limit yourself via a setting or configuration file \u2014 it depends on the model you're using and your account\u2019s tier with Anthropic. To access a larger context window, there are a few specific steps and requirements. Here's how it works and what you need to do:\n\n---\n\n## \ud83d\udd0d What are the existing limits\n\n- Most Claude API models (e.g. Claude Sonnet 4 or Opus 4.1) have a **200,000-token context window** by default. ([datastudios.org](https://www.datastudios.org/post/claude-context-window-token-limits-memory-policy-and-2025-rules?utm_source=openai))  \n- Using these models, the maximum number of **output tokens** (i.e. the result you get back) is limited (for example, standard limit is ~64,000 output tokens for Sonnet 4). ([datastudios.org](https://www.datastudios.org/post/claude-context-window-token-limits-memory-policy-and-2025-rules?utm_source=openai))  \n- For higher-end or enterprise use, there are options for much larger context windows. For instance, in Claude Code (a special mode), certain paid plans can enable up to **1 million tokens** of context. ([support.claude.com](https://support.claude.com/en/articles/8606394-how-large-is-the-context-window-on-paid-claude-plans?utm_source=openai))  \n\n---\n\n## \ud83d\udee0 How you can \u201cget\u201d a larger context window\n\nSince you can\u2019t directly tweak the limit yourself, here\u2019s what you can do to upgrade or request higher capacity:\n\n| Step | What to do |\n|---|---|\n| 1 | **Check your current plan/tier** in the Anthropic Console (usually under *Billing* or *Settings \u203a Plans*). Your token limits are tied to the tier. ([claudeapipricing.com](https://claudeapipricing.com/rate-limits?utm_source=openai)) |\n| 2 | If you're using one of the **paid Claude plans** (such as Pro, Team, Enterprise) and need more than the standard limits, **see if there\u2019s an \u201cExtra Usage\u201d (or \u201cEnable higher context\u201d) toggle in Claude Code**. That often unlocks 1M token contexts. ([support.claude.com](https://support.claude.com/en/articles/8606394-how-large-is-the-context-window-on-paid-claude-plans?utm_source=openai)) |\n| 3 | Otherwise, go to **Anthropic\u2019s support or sales contact** (via the Console) and request extended limits. For enterprise-level usage, they offer **custom context windows and output-token limits**. ([docs.anthropic.com](https://docs.anthropic.com/en/api/rate-limits?utm_source=openai)) |\n| 4 | In your API code, make sure you\u2019re using a model that supports the larger limit (e.g. Opus 4.6, Sonnet 4.6, etc.)\u2014some models in earlier or \u201cHaiku\u201d versions have lower limits. Also check if the required beta or \u201cextra usage\u201d headers or settings are applied. ([datastudios.org](https://www.datastudios.org/post/claude-context-window-token-limits-memory-policy-and-2025-rules?utm_source=openai)) |\n\n---\n\nIf you tell me which model you're using and what tier/account type you have, I can help you figure out exactly whether you *already* have access to larger limits or whether you need to request them. Do you want me to check for **your specific account**?"
    }
]