Price checkPer million tokens
Trial (Free)FreePay-as-you-go (Legacy Models)Usage-basedModel Vault (Dedicated Instances)Custom
See all 4 plans

All Cohere API Plans & Pricing

Plan Monthly Annual Best For
View all features by plan (compare side-by-side)

Trial (Free)

  • Free trial API key (auto-created on signup)
  • Rate-limited access to all models
  • Not permitted for commercial or production use
  • No credit card required

Pay-as-you-go (Legacy Models)

  • Command: $1.00 input / $2.00 output per M tokens
  • Command-light: $0.30 input / $0.60 output per M tokens
  • Command R (03-2024): $0.50 input / $1.50 output per M tokens
  • Command R+ (04-2024): $3.00 input / $15.00 output per M tokens
  • Command R+ (08-2024): $2.50 input / $10.00 output per M tokens
  • Aya Expanse 8B/32B: $0.50 input / $1.50 output per M tokens
  • Production API key required (apply via dashboard)
  • Billed monthly or at $250 outstanding balance

Model Vault (Dedicated Instances)

  • Embed 4 Small: $4.00/hr or $2,500/mo per instance
  • Embed 4 Medium: $5.00/hr or $3,250/mo per instance
  • Rerank 3.5 Medium: $5.00/hr or $3,250/mo per instance
  • Rerank 4 Fast Medium: $5.00/hr or $3,250/mo per instance
  • Rerank 4 Pro Medium: $5.00/hr or $3,250/mo per instance
  • Rerank 4 Pro Large: $10.00/hr or $6,500/mo per instance
  • Fully managed, no shared resources or multi-tenancy
  • Fixed or Flex pricing plans available
  • Annual commitment options available

Enterprise (Current Models)

  • Current Command generation (Command A, new Command): contact sales
  • North AI platform: custom enterprise pricing
  • Compass intelligent search: custom enterprise pricing
  • North Mini Code (agentic coding): contact sales
  • Private deployments available
  • Custom model fine-tuning and bespoke customization
Pricing Alerts

Track Cohere API pricing

Get an email when Cohere API's pricing changes — plus the weekly SaaS Price Watch: verified price changes and deals across 3,000+ products. One-click unsubscribe.

Start Cohere API Free Trial
Compare Cohere API with alternativesAdjust seats, lock a tier, add up to 2 more products side-by-side. Shareable URL.
Quick Answer
Last verified:
Estimate

Cohere API offers usage-based pricing from $0.300–$15.00 per million tokens as of August 2026 and custom pricing for larger requirements. Plans: Trial (Free) (free), and Pay-as-you-go (Legacy Models) (usage-based). Custom pricing is available on request. Usage cost depends on the selected model and generation volume.

Use the interactive pricing calculator to estimate your exact cost based on team size and requirements.

  • Free tier: Yes

Cohere API offers 4 pricing tiers: Trial (Free), Pay-as-you-go (Legacy Models), Model Vault (Dedicated Instances), Enterprise (Current Models). A free plan is available. Paid plans include Pay-as-you-go (Legacy Models) (usage-based). The Pay-as-you-go (Legacy Models) plan is existing customers on legacy command model contracts.

Compared to other llm api providers software, Cohere API is positioned at the budget-friendly price point.

  • 1 documented hidden costs beyond list price
Last verified: | Sources: 1 pricing pages and reviews

How much does Cohere API cost?

Cohere API offers usage-based pricing from $0.300–$15.00 per million tokens, with custom pricing available for larger requirements. Plans include Trial (Free) (free), Pay-as-you-go (Legacy Models) (usage-based), Model Vault (Dedicated Instances) (custom pricing), Enterprise (Current Models) (custom pricing).

Cohere API Pricing Overview

Cohere API offers usage-based pricing from $0.300–$15.00 per million tokens and custom pricing for larger requirements. The Trial (Free) plan is free and is best for evaluation and prototyping. The Pay-as-you-go (Legacy Models) plan is usage-based and is designed for existing customers on legacy command model contracts. The Model Vault (Dedicated Instances) plan requires contacting sales for a custom quote and is designed for teams needing dedicated, private, guaranteed-performance inference at scale. The Enterprise (Current Models) plan requires contacting sales for a custom quote and is designed for enterprises requiring current-generation models, private deployments, or ai platform access.

There are at least 1 documented hidden costs beyond Cohere API's list price, including implementation, training, and add-on fees.

This pricing was last verified in July 27, 2026 from 1 independent source.

Start Cohere API Free Trial
Cohere is purpose-built for enterprise retrieval-augmented generation (RAG). The Command R7B model starts at just $0.037 per million input tokens, making it the most affordable in Cohere's lineup for high-volume RAG tasks. Command R runs $0.15/$0.60 per million tokens with native grounding and citation. Command R+ and Command A cost $2.50/$10.00 per million tokens for complex agentic workflows. Cohere's Embed and Rerank APIs round out a complete RAG stack — Embed v3 at $0.10/M tokens and Rerank 3.5 at $2.00 per 1,000 queries. A free Trial key is available for non-commercial evaluation.

How Cohere API Pricing Compares

Compare Cohere API pricing against top alternatives in LLM API Providers.

OpenAI $0-$200/month Compare → Groq $0-$3.0 per million tokens Compare → Together AI $0.03-$9.95 per million tokens / hour Compare →

Usage-Based Rates

Per-unit pricing for Cohere API API usage.

Pay-as-you-go (Legacy Models)

Model Input Output Cached Per
command $1.00 $2.00 — 1M tokens
command-light $0.300 $0.600 — 1M tokens
command-r-03-2024 $0.500 $1.50 — 1M tokens
command-r-plus-04-2024 $3.00 $15.00 — 1M tokens
command-r-plus-08-2024 $2.50 $10.00 — 1M tokens
aya-expanse-8b $0.500 $1.50 — 1M tokens
aya-expanse-32b $0.500 $1.50 — 1M tokens
  • Labeled as legacy pricing for existing customers on the live page
  • Current-generation models (Command A, new Command, North Mini Code) require enterprise contact
  • Command R7B and Command R PAYG prices ($0.037/$0.15, $0.15/$0.60) no longer listed publicly

Compare Cohere API vs Alternatives

Before committing to Cohere API, compare pricing with these 3 alternatives in the same category.

VSOpenAI
Free

Casual users exploring AI for everyday tasks

Full comparison
VSGroq
Free

Prototyping and evaluation

Full comparison
VSTogether AI
From $0.03 per million tokens / hour

Variable-volume API usage

Full comparison

All Cohere API alternatives & migration guides

What Companies Actually Pay for Cohere API

Median per-1M-token pricing across 1 models
Input $2.50/1M
Output $10.00/1M
Flagship models in this provider's catalog
Model Input /1M Output /1M Blended /1M
cohere_command-a $2.50 $10.00 $4.38
Review scores
Third-party review aggregates, as of Apr 2026
Top pricing complaints
Price-to-performance ratio considered uncompetitive versus GPT-4o and Claude SonnetPricing system described as disappointing after the free trial tier endsPricing page difficult to find — not accessible from main navigation
Source: Artificial Analysis — medians aggregated from 1 models in this provider's catalog. Per-1M-token pricing reflects list rates.

How Cohere API Pricing Compares

Software Starting Price Top Price
Cohere API Free $15 per million tokens
Amazon Bedrock $0.07 per million tokens $75 per million tokens
Anyscale Free $4.9591 per million tokens
Baidu ERNIE API $0.1 per million tokens $10 per million tokens
Cerebras Inference API $0.1 per million tokens $6 per million tokens
Claude API $0.1 per million tokens $75 per million tokens

Detailed pricing comparisons:

Browse all LLM API Providers pricing →

1 Cohere API Hidden Costs Beyond the List Price

Beyond the listed price, Cohere API has at least 1 documented hidden costs that can significantly increase total cost of ownership.

Watch for 1 hidden costs
  • Pricing Jump After Free Trial Tier 5-15% of license costs
    medium 1 source
    Reddit "I've also seen some decent results by using Cohere's Command R Plus + Cohere reranker but their pricing system is a bit of a disappointment after the free tier."
Tip

Ask your Cohere API sales rep about these costs upfront. Getting them in writing before signing can save you from surprise charges later.

Full hidden costs breakdown →

Intelligence sourced from 1 independent sources
Reddit User discussions
Key claims include inline source attribution. Data verified against multiple independent sources. 2 source citations total.

Cohere API Contract Terms

Cohere API contracts do not auto-renew. Changes require advance notice. These terms are sourced from verified buyer experiences.

Contract Terms
Auto-Renewal No
Mid-Term Downgrade Allowed
Payment Terms Pay-as-you-go, billed per token consumed
Note

Pay-as-you-go model — usage can be scaled down or stopped at any time with no commitment

Cohere API Price History

Pricing changes CostBench has tracked for Cohere API, verified across 2 snapshots going back to 2026Q2.

  1. Cohere API tier renamed

    "Command R (Pay-as-you-go)" renamed to "Pay-as-you-go (Legacy Models)"

  2. Cohere API tier renamed

    "Command R+ / Command A (Pay-as-you-go)" renamed to "Model Vault (Dedicated Instances)"

  3. Cohere API tier renamed

    "Embed & Rerank" renamed to "Enterprise (Current Models)"

View Cohere API's full price history →

Cohere API Pricing FAQ

01 How much does the Cohere API cost?

Cohere API pricing varies by model type. Command R7B (the cheapest generative model) costs $0.037 per million input tokens and $0.15 per million output tokens. Command R costs $0.15/$0.60 per million tokens. Command R+ and Command A cost $2.50/$10.00 per million tokens. Embed v3 costs $0.10 per million tokens. Rerank 3.5 costs $2.00 per 1,000 queries.

02 What is Command R and when should I use it?

Command R is Cohere's retrieval-augmented generation (RAG) model, optimized for grounded responses with cited sources. Use Command R ($0.15/$0.60 per million tokens) for everyday RAG pipelines. Use Command R+ or Command A ($2.50/$10.00 per million tokens) when you need higher accuracy on complex multi-step tasks. Command A features a 256K context window — the largest in Cohere's lineup.

03 What is Cohere Rerank and how is it priced?

Cohere Rerank re-orders search results by relevance using a cross-encoder model, dramatically improving RAG retrieval quality. Rerank 3.5 costs $2.00 per 1,000 queries. For high-volume applications, Cohere also offers hourly ($5/hr) or monthly ($3,250/mo) dedicated Model Vault capacity. Rerank is often used alongside Embed in a two-stage retrieval pipeline.

04 Does Cohere have a free tier?

Yes, Cohere offers a free Trial API key with rate-limited access to all models. However, the Trial key is explicitly not permitted for commercial or production use. For production applications, you need a Production API key, which uses pay-as-you-go billing charged monthly or when you reach $250 in outstanding usage.

05 Cohere vs OpenAI for enterprise RAG: which is better?

Cohere is purpose-built for enterprise RAG and search use cases, while OpenAI is more general-purpose. Cohere's Command R models include native grounding, citation generation, and multi-step tool use optimized for RAG pipelines. Cohere's Embed and Rerank models are best-in-class for retrieval. For pure RAG performance, Cohere often wins. For general-purpose chat or coding tasks, OpenAI or Anthropic may be better choices.

06 What does Cohere API actually cost per token?

Based on Artificial Analysis data (April 2026), Cohere's Command A model is priced at $2.50 per 1M input tokens and $10.00 per 1M output tokens, for a blended rate of approximately $4.38/1M tokens. Command R and Embed & Rerank tiers are custom pay-as-you-go pricing — check cohere.com/pricing for current rates.

07 How does Cohere pricing compare to OpenAI and Anthropic?

Community feedback from 2024 indicates Cohere's Command R+ pricing was viewed as uncompetitive relative to GPT-4o and Claude Sonnet 3.5, with some teams switching providers. Cohere's Command A is currently priced at $2.50/1M input and $10.00/1M output tokens per Artificial Analysis — compare directly against current OpenAI and Anthropic rates when evaluating.

08 Is there a minimum commitment for the Cohere API?

No. Cohere uses a pay-as-you-go model with no minimum commitment or annual contract required. You pay only for the tokens you consume across the Command R, Command R+ / Command A, and Embed & Rerank tiers.

Is this pricing incorrect? — we'll verify and update it.