claude-opus-4-5 — pricing, context window and limits
| Status | legacy |
|---|---|
| Input / MTok | $5 |
| Output / MTok | $25 |
| Context window | 200K tokens |
| Max output | 64K tokens |
| Min cacheable prefix | 4,096 tokens |
| Effort levels | low, medium, high |
| Thinking | budget-tokens |
| Platforms | Claude API |
Still active, and the first Opus with effort control at all
Opus 4.5 was the generation that introduced the effort parameter to the Opus line, though only
at three levels rather than the fuller range later generations support — both the highest tier and
the one below it are unavailable here and error if requested. It remains an active, supported
model rather than a deprecated one, distinct from the older, genuinely retiring generations further
back in the lineup.
Where this sits relative to the current top tier
Compared to Opus 5, this generation carries the older, opt-in-only thinking behaviour and a narrower effort range, at a matching sticker price — which makes the newer generation's expanded range close to a straightforward upgrade for teams already comfortable spending at this tier.
Why teams still on this generation tend to stay a while
Because it's still fully active rather than deprecated, there's no forcing function pushing a migration off this model the way there is for a genuinely retiring one — which means it's worth periodically re-asking whether the newer generation's wider effort range and different thinking default would actually change your outcomes, rather than letting "it still works" substitute for a deliberate decision to stay.
Related
See what actually changed from Opus 4.6 to Opus 5 for the fuller picture of what's changed since this generation, even though that guide is framed against 4.6 specifically.
Verified 2026-08-08 against claude-api skill — shared/models.md + SKILL.md model table (canonical: https://platform.claude.com/docs/en/about-claude/models/overview, https://platform.claude.com/docs/en/pricing).Could not confirm: Context window and max output for this legacy model are not restated in our source; the figures shown are the documented defaults for its generation and should be re-checked against the Models API before you budget against them.