Anthropic: Claude Opus 4.7

anthropic/claude-opus-4.7-fast is where the fast-mode option makes its debut in this line, heading at not publishedlast known value, as of an unrecorded date — the source no longer lists this rowopenrouter-models, last checked 2026-09-14 input against this row's own $5.00per million tokensopenrouter-models, last checked 2026-09-14 — six times over. In the next release the ratio fell by two thirds: anthropic/claude-opus-4.8-fast lists at not publishedlast known value, as of an unrecorded date — the source no longer lists this rowopenrouter-models, last checked 2026-09-14 against $5.00per million tokensopenrouter-models, last checked 2026-09-14, twice over rather than six, and Opus 5 kept that same rate (not publishedlast known value, as of an unrecorded date — the source no longer lists this rowopenrouter-models, last checked 2026-09-14 against $5.00per million tokensopenrouter-models, last checked 2026-09-14). Each of those six figures is the top listed provider's rate for its row rather than necessarily Anthropic's own, and two rows are not obliged to be headed by the same provider, so these are the multipliers between listings as the catalog stands in the snapshot of 28 August 2026.

The six-times option had a short life. anthropic/claude-opus-4.7-fast was listed on 12 May 2026 and anthropic/claude-opus-4.8-fast replaced it fifteen days later — and nothing else about the option is recorded as having changed across that handover: same context window, same base rate, only the multiplier sitting on top of it. Once it dropped, it stayed dropped, and Opus 5's fast row still lists at the two-times rate one full generation later.

This row's base price is itself part of a longer plateau — $5.00per million tokensopenrouter-models, last checked 2026-09-14, unchanged from anthropic/claude-opus-4.5 through anthropic/claude-opus-5, with the batch rate alongside it flat across the same five releases. So among the prices in this family the fast-mode cut is the only one that moved. Other things did move in that stretch — the context window at 4.6, and the Artificial Analysis intelligence index at each release after 4.7 — which is what makes the flat price line worth pointing at rather than assuming.