Cohere

Command A API Pricing

Current flagship Command model; per-token pricing not published on cohere.com/pricing as of verification date (self-serve pricing appears to have moved behind Model Vault / enterprise sales for this generation).

Could not verify pricing on 2026-07-02 — treat these numbers as tentative.

View source ↗

Facts

Unverified

Specs

  • FamilyCommand A
  • API IDcommand-a-03-2025
  • ModalityText
  • Context window256K tokens
  • Max output8K tokens
  • Release date2025-03-01

Pricing (per 1M tokens)

  • Input
  • Output
  • Cached input
  • Cache write
  • Batch discount

Notes

Current flagship Command model; per-token pricing not published on cohere.com/pricing as of verification date (self-serve pricing appears to have moved behind Model Vault / enterprise sales for this generation).

Worked cost examples

Not priced per token

Command Ais billed per unit (image, second, or similar), not per token — the standard chatbot/RAG/batch workloads don’t apply. See the notes in the facts panel above for its per-unit rate.

Price history

Changes we have documented since we began tracking

No price changes recorded since we began tracking.

Command A pricing FAQ

How is Command A priced?
Command A is billed per unit rather than per token. Current flagship Command model; per-token pricing not published on cohere.com/pricing as of verification date (self-serve pricing appears to have moved behind Model Vault / enterprise sales for this generation).
Does Command A have cache pricing?
No cache discount is documented for Command A on the provider's pricing page.