muse-spark-1.1

A Meta (Llama) model. Every figure below was read off Meta (Llama)'s own pricing page and stamped with the date it was read — last pass 2026-08-04. Nothing here is estimated.

Input / 1M tokens$1.25
Output / 1M tokens$4.25
Released2026-07-09

Meta's FIRST paid first-party API (agentic coder). cached_input is community-reported, not vendor-confirmed

vendor-published as of 2026-08-04: ai.developer.meta.com/docs/pricing-rate-limits.md — $1.25/$4.25, cached input $0.15

Price is the easy half. What it costs you also depends on how fast the first token arrives and what survives concurrency — that part is measured. See every Meta (Llama) model on its provider page.