LLM API

Cohere API pricing

Compare source-tracked Cohere Command A, Command R, and Command R+ API pricing, context limits, and snapshot-specific price changes.

Decision summary: Command A is Cohere's current agent-oriented model with a 256K context window. Command R remains the lower-cost tracked option, while Command R+ and Command A share the same recorded token rate but differ in model generation and specifications.
Pricing checked 2026-07-17high confidence

Pricing overview

Cohere bills hosted production API usage by input and output tokens. Product names and timestamped API model IDs should both be verified before deployment.

Source-tracked model data

Cohere API model prices in one table

Example cost uses 1,000,000 input tokens and 300,000 output tokens at each model's default tracked rate, with no cache discount.

ModelInput priceCached inputOutput priceContextExample cost
Command R 08-2024
Stable
$0.15 / 1MNot tracked$0.60 / 1M128K tokens$0.33
Command R+ 08-2024
Stable
$2.50 / 1MNot tracked$10.00 / 1M128K tokens$5.50
Command A
Stable
$2.50 / 1MNot tracked$10.00 / 1M256K tokens$5.50

Cost boundary: this example compares token charges only. It does not assume equal output quality, latency, retry rates, or output length across models.

Official price history

Cohere API recorded model price changes

Verified changes for exact models and pricing profiles available through this provider. New model generations are not treated as price changes.

Cohere · Standard

Command R 08-2024

Effective 2024-08-30
RateEarlier priceLater recorded priceChangePercentage
Input / 1M$0.50$0.15-$0.35-70.0%
Output / 1M$1.50$0.60-$0.90-60.0%
Checked 2026-07-17Official source ↗

API snapshot: command-r-03-2024command-r-08-2024

Cohere released the Command R 08-2024 API snapshot at lower hosted-API rates than the March snapshot.

Cohere · Standard

Command R+ 08-2024

Effective 2024-08-30
RateEarlier priceLater recorded priceChangePercentage
Input / 1M$3.00$2.50-$0.50-16.7%
Output / 1M$15.00$10.00-$5.00-33.3%
Checked 2026-07-17Official source ↗

API snapshot: command-r-plus-04-2024command-r-plus-08-2024

Cohere released the Command R+ 08-2024 API snapshot at lower hosted-API rates than the April snapshot.

What affects cost

  • Selected Command API snapshot
  • Input and output token volume
  • Output length and retry behavior
  • Choice between Command R and Command R+
  • Production API versus private deployment terms

Lower-cost options from the same provider

  • Command R 08-2024 instead of Command R+ when workflow tests show it is sufficient
  • Shorter outputs and fewer retries

Alternative providers or products

  • Google Gemini API
  • Anthropic API
  • OpenAI API
  • Mistral API

Best for

  • enterprise retrieval-augmented generation
  • workflows that need generated citations
  • teams using Cohere tool-use and agent APIs

Not ideal for

  • teams that do not pin an exact API snapshot
  • buyers treating private deployment pricing as token pricing
  • new workloads that have not evaluated Cohere's newer Command A family
FAQ

Cohere API pricing questions

How much does Command A cost?

Cohere lists Command A at $2.50 per 1M input tokens and $10 per 1M output tokens. Its official page currently documents the command-a-plus-05-2026 API model ID.

How much does Command R 08-2024 cost?

Cohere lists Command R 08-2024 at $0.15 per 1M input tokens and $0.60 per 1M output tokens on its hosted API.

Are Command R 03-2024 and Command R 08-2024 the same API model ID?

No. Cohere uses timestamped API model IDs. StackLens shows the transition explicitly and does not hide the snapshot change behind a generic model name.

Does Cohere recommend Command R for every new project?

No. Cohere's current model documentation recommends evaluating Command A for most new use cases while continuing to list the August 2024 Command R models as live.