Command (full Command text / Command A‑class)

Chat / reasoning / coding

Typically 128K–256K tokens (check Cohere / Orq docs for current limits)

Complex reasoning, analysis, coding, and high‑stakes tasks where you want Cohere’s strongest general‑purpose Command model

Premium – Cohere standard API pricing; current Command‑class models are around 1.00 USD / 1M input tokens and 2.00 USD / 1M output tokens for text, with higher tiers for advanced variants

Command R

Chat / reasoning / RAG

Up to 128K tokens (verify in Cohere / Orq docs)

Everyday production workloads, retrieval‑augmented generation, coding, product features, and workflows that balance quality with cost and latency

Mid‑tier – Command R pricing around 0.15 USD / 1M input tokens and 0.60 USD / 1M output tokens (cache‑miss rates), making it a strong option

Command R7b / Command Light

Chat / fast / cost‑efficient

Up to 128K tokens (confirm per model), smaller parameter count

Fast, lower‑cost tasks such as classification, extraction, lightweight chat, routing, and high‑volume support workflows where 7B‑class models are enough

Cost‑efficient – Command Light and Command R7b models can be as low as ~0.03–0.30 USD / 1M input tokens and ~0.15–0.60 USD / 1M output tokens depending on the specific variant

Chat

Cohere Command family (Command, Command R, Command Light/R7b)

Call Command models directly via Cohere’s API for chat, reasoning, coding, and RAG use cases

Use Cohere models through Orq.ai’s OpenAI‑compatible endpoint, adding routing, tracing, evals, budgets, and governance controls around each request

Code

Command / Command R models tuned or suitable for coding

Use Cohere directly for code generation, debugging, refactoring, and agentic coding workflows

Route coding workloads through Orq.ai, compare Cohere Command models against other providers, and monitor cost, latency, and quality from one control layer

Embeddings & rerank

Embed and Rerank models configured in Orq.ai

Cohere offers specialized Embed and Rerank models for search, ranking, and retrieval

Use Orq.ai to route embedding and rerank workloads to those specialized models while keeping Command/Command R for reasoning, generation, and agent steps

Model rates

Cohere model pricing may differ depending on whether you:

* connect your own Cohere account and API key (BYOK), or

* use Cohere models billed through Orq.ai where available Check the Orq.ai pricing page and your workspace’s provider configuration for current per‑model rates, quotas, and plan details.