claude-opus-4-7 — pricing, context window and limits

The generation that introduced the current tokenizer

Opus 4.7 is where the tokenizer the 4.8 and 5 generation still uses was introduced, along with the highest effort tier below the absolute maximum and high-resolution vision support — real, foundational capability the generation before it didn't have. Anything built against this model or its successors inherits that tokenizer's counting behaviour, which is worth knowing if you're comparing token counts against an older Opus generation.

The cache-minimum outlier

This model's minimum cacheable prefix is higher than both the generation immediately before it and the generation immediately after — the single most expensive thing to assume incorrectly about this specific model if you're used to reasoning about cache minimums as trending steadily in one direction across generations. See the minimum-cacheable-prefix table for the full, non-monotonic picture across the whole lineup.

Fast mode was removed here specifically

Unlike the Opus generation before and after it, this one does not support fast mode at all — requesting it is rejected. See speed: "fast" is not supported on Opus 4.7 for exactly what that failure looks like and how to route around it.

Related

See migrating: Opus 4.7 to Opus 5 for the field-by-field diff against the current top-tier successor.