Frontier builder · United States
Anthropic
Frontier IC25 · Edition 2026.3 · Evidence reviewed October 8, 2026 UTC
A model builder whose Claude APIs expose reasoning, tool use and coding workflows.
Selected for a consequential, inspectable general-purpose model contribution. The displayed model is a named representative, not a claim of latest or best performance.
Official transparency page explicitly identifies US headquarters.
Selected model
Claude Sonnet 5.5 · claude-sonnet-5-5
Standard first-party synchronous API; uncached text tokens, prompt ≤100K.
1,000,000 · Context window
Scope and limitations
- Adaptive thinking is enabled by default; thinking tokens are billed as output and count toward output limits, even when returned thinking is omitted or summarized.
- 1 million context tokens are available at standard price; the legacy >200K premium does not apply to this selected model.
- US-only inference has a 1.1x token-price multiplier. Partner platforms have independent regional pricing.
- The 300K output beta is Batch API only and is excluded from this synchronous comparison.
Original sources
Explore the full IC25 tool · Read the field guide in six languages