Context windows › Claude

Claude context window: 1,000,000 tokens

The current Claude models from Anthropic share one context window size. Fable 5, Opus 5, Opus 4.8, Sonnet 5, and Sonnet 4.6 all have a 1,000,000-token context window with a 128,000-token max output. The exception is Haiku 4.5, the small model, at 200,000 tokens of context and 64,000 max output. If you searched for the Claude context window size, the short answer for every current full-size model is one million tokens.

Verified 2026-08-10.

No long-context surcharge

On the current Claude models, the 1M context window is available at standard pricing. There is no long-context surcharge: a 900,000-token prompt is billed at the same per-token rate as a 9,000-token prompt. That contrasts with Gemini 2.5 Pro, where the per-token price rises once a prompt passes 200,000 input tokens. The bill still scales with tokens sent, since input billing always does, but the rate stays flat across the whole window. Current rates are on the Claude pricing page.

Searching for the Claude 4 context window?

A lot of queries for this page ask about older Claude 4-era models. Those models have been superseded by the current lineup, and the current lineup is the one in the table above: 1,000,000 tokens of context on every full-size model, 200,000 on Haiku 4.5. If you are choosing a model today, the older figures no longer apply to anything you can pick.

Using the full million

Two practical notes at this scale. First, cost ceiling: a window this size means one request can carry up to a million billable input tokens, so budget per request, not just per token. Second, quality: long prompts can degrade retrieval of facts placed in the middle of the window, an effect well documented in long-context evaluations. Keep the key material near the start or end of the prompt and test at your real prompt length.

More on the model family is on the Claude overview. For every current model side by side, see the context window comparison.