Retell AI Pricing in 2026: Plans, Per-Minute Costs, and What You Actually Pay 

Retell AI is a developer-first platform for building AI voice agents. It advertises a simple rate starting at $0.07/min, but that’s only one layer of the total Retell AI cost. This guide breaks down what Retell AI pricing plans include, what Retell AI costs per minute in practice, and where the hidden costs show up.

Retell AI pricing at a glance

Retell AI offers two true pay-as-you-go pricing tiers:

• Pay as you go: $0.07-$0.31/min for voice agents, $0.002+/msg for chat.
• Enterprise: Custom pricing for both voice agents and chat.

Note: Retell AI pricing is modular: your final rate is the sum of voice infrastructure, text-to-speech, LLM, and telephony, each billed separately.

What does the Retell AI pricing plan offer

Retell AI uses two main pricing options: Pay as you go for usage-based billing and Enterprise for organizations with larger volumes and specific business requirements. Both plans provide access to the core Retell AI platform, while Enterprise adds dedicated infrastructure, higher concurrency, custom terms, and dedicated support.

Pay-as-you-go plan

Retell AI’s Pay-as-you-go plan is its usage-based option for teams that want to build, test, and deploy AI voice agents without a fixed monthly subscription or long-term contract. You get full platform access, start at $0, and pay based on the minutes and services you use. 

New accounts also receive $10 in free credits, giving users room to test their agents before incurring regular usage charges.

How much does it cost?

The starting platform range for Pay-as-you-go is $0.07 to $0.31 per minute. For AI chat agents, you need to pay $0.002+ per minute. These prices aren’t a single fixed rate, so your actual Retell AI cost per minute depends on the LLM, text-to-speech (TTS), telephony provider, and additional features attached to your agents.

What’s included in Retell AI’s Pay-as-you-go plan

The Pay-as-you-go plan provides access to the core Retell AI platform and its main agent-building and monitoring tools. The included features are:

• Full platform access: Build and deploy AI voice agents without feature restrictions
• Pre-built agent templates: Start agent development with ready-made templates
• Call analytics and transcripts: Review call activity and conversation records
• Simulation testing: Test agents before putting them into production
• Webhooks and API access: Connect agents with external applications and workflows
• 20 concurrent calls: Each account gets 20 concurrent calls at no additional charge
• Community and email support: Get help through Retell AI’s standard support channels

The plan also follows a usage-based billing model, so there are no minimum usage commitments. You can purchase additional concurrent call capacity as your call volume grows.

What’s only available as an add-on 

Several services are priced separately from the core per-minute rate. These can increase your total Retell AI fees, depending on how you configure your agent.

The underlying LLM, voice, and telephony services can also affect the final bill as their rates vary by provider and model. This means the advertised Retell AI cost per minute should be treated as a starting point. A production agent using a higher-priced LLM, premium voice, telephony, and paid add-ons can have a higher effective per-minute cost.

Who this model works best for

Pay-as-you-go works for developers, startups, growing businesses, and teams testing AI phone agents. It suits projects where call volume changes from month to month. 

It is also useful when you want to start with a small deployment and gradually increase usage without committing to an Enterprise contract. The $10 free credits let you test agent behavior, call flows, and integrations before paying for regular usage.

Enterprise plan

Retell AI’s Enterprise plan is a custom plan designed for organizations running larger AI voice deployments that need higher call capacity, dedicated infrastructure, custom contractual terms, and dedicated support. It includes the capabilities available in Pay-as-you-go and adds features for larger operational and compliance requirements.

How much does it cost?

Retell AI Enterprise pricing is custom. Organizations need to contact the sales team to discuss their requirements and receive custom pricing. The final cost can depend on factors such as call volume, concurrency requirements, infrastructure needs, compliance requirements, and the services included in the agreement.

What’s included in Retell AI’s Enterprise plan

Enterprise includes the Pay-as-you-go features plus additional capabilities for larger deployments:

• Dedicated stable server: Provides dedicated infrastructure for Enterprise workloads
• High concurrent-call capacity: Enterprise accounts have a higher concurrency limit, with custom capacity available based on requirements
• Custom MSA and DPA terms: Organizations can establish contractual and data-processing terms suited to their requirements
• Role-based access control: Manage user permissions across the organization
• Custom SSO: Support for customized single sign-on requirements
• 24/7 support: Enterprise customers receive round-the-clock support through a dedicated portal
• Dedicated implementation support: Available as an add-on for organizations that need additional deployment assistance.

Who this model works best for

The Enterprise plan suits large organizations, high-volume contact operations, and businesses with strict security or compliance requirements. It makes sense when standard concurrency limits, shared infrastructure, or shared support no longer fit the deployment.

For teams handling many simultaneous calls, Enterprise provides custom concurrency and dedicated infrastructure. It also fits organizations that require formal contracts, role-based access, SSO, dedicated support, or specific compliance arrangements.

Retell AI Pay-as-you-go vs. Retell AI Enterprise plan

Retell AI offers the same core platform across both plans, while pricing, infrastructure, call capacity, and support differ.

Retell AI tracks calls to the nearest second and calculates billing from your total accumulated usage. You are charged for the full duration of a connected call, including periods when the agent is listening during silence or hold time. Failed calls that never connect are not billed.

What makes up Retell’s per-minute cost

The advertised Retell AI cost per minute covers only a piece of the stack. A working voice agent stacks four components, and each is billed separately.

• Retell Voice Infrastructure ($0.055/min):

This base platform layer runs your AI agents and handles voice interaction. It applies to every call regardless of which TTS provider or LLM you choose.

• Text-to-Speech Costs:

Text-to-speech converts the AI agent’s responses into spoken audio. Retell lists $0.015/min for providers such as Retell, Minimax, Fish, Cartesia, and OpenAI. ElevenLabs costs $0.04/min, so your voice provider choice is a major factor in total Retell AI fees.

• LLM Costs:

The LLM generates the agent’s responses during a call. Pricing varies significantly by model, ranging from around $0.003/min for GPT-5 nano to $0.16/min for GPT-5.5. Retell also applies a Fast tier surcharge when you select faster processing, which can increase the LLM portion of your bill.

• Telephony Costs:

Telephony charges cover the phone network connection used for calls. Retell lists $0.015/min when using providers such as Twilio or Telnyx. Custom telephony is free, so businesses using their own telephony setup can avoid this additional per-minute charge.

Additional or hidden Retell AI costs to budget for

Beyond the four core layers, several add-ons and overlooked billing details push the real Retell AI cost higher than the base rate suggests.

• Knowledge Base overage: +$0.005/min; 10 Knowledge Bases are free; additional ones cost $8/month each
• Hold time and silence: calls are billed for the full connected duration, including hold music and dead air
• LLM upgrades: Choosing a higher-priced LLM or Fast Tier increases the LLM portion of your bill by 1.5 to 2x. For example, GPT-5.5 costs $0.16/minute on the standard tier and $0.32/minute on the Fast Tier.
• Premium voices: ElevenLabs costs $0.04/minute compared with $0.015/minute for other listed voice providers.
• Long prompts: LLM billing can increase when prompts exceed the applicable token threshold, so large system prompts and conversation context can affect your Retell AI billing.
• Enterprise services: Enterprise customers can have custom pricing for higher usage and requirements.

What counts as a billable minute on Retell AI?

Retell AI billing is based on the actual time your voice agent spends handling a connected call. Understanding what counts towards usage helps you estimate your Retell AI cost per minute more accurately.

Here’s what counts as a billable minute:

• Connected calls: Once a call connects, the conversation time counts toward your usage.
• Agent listening time: Time spent listening to the caller, including periods of silence during an active call, is included.
• Agent speaking time: Time the AI voice agent spends responding to the caller is billable.
• Hold or waiting time: Time spent waiting during a connected call also counts toward the call duration.
• Inbound and outbound calls: Both types of connected calls contribute to your usage.
• Call duration: Retell tracks call duration to the nearest second, so billing is based on actual usage rather than rounding every call to a full minute.

Here’s what doesn’t count as a billable minute:

• Calls that never connect: Unanswered or failed calls that don’t establish a connection aren’t billed as call minutes.
• Unused account credits: Free or purchased Retell AI credits are consumed only when you generate billable usage.
• Time before a call connects: Ringing time before the call is answered does not count as connected-call duration.

Retell AI billing exception

• Calls shorter than 10 seconds that use a dynamic opening message (AI speaks first) carry a minimum 10-second charge, since generating that opening line requires processing time regardless of call length.
• Agents using prompts over 4000 LLM tokens have their billed duration scaled proportionally, since longer prompts add processing overhead per call.

Retell AI pricing example: Calculate your monthly bill

A simple way to estimate your monthly Retell AI cost is to add core voice infrastructure, model and voice charges, telephony, and any other services your AI agent uses.

Formula:

Monthly Retell AI cost = AI agent usage + model/voice costs + telephony costs + other applicable charges

The examples below use hypothetical usage and the listed Retell rates to show how monthly costs can vary by use case.

1. Customer support agent

A support agent handling FAQs and basic troubleshooting can run on a lighter stack.

• AI agent usage: 1000 min
• Retell voice infrastructure: $0.055/min
• TTS (Retell platform voices): $0.015/min
• LLM (GPT-5 Mini): $0.012/min
• Telephony (Retell Twilio): $0.015/min

Monthly Retell AI cost = 1000 x ($0.055+$0.015+$0.012+$0.015) = $97/month

2. Appointment scheduling agent

Scheduling calls are short and structured, so a mid-range LLM with Knowledge Base support for availability lookups is typical.

• AI agent usage: 2000 min
• Retell voice infrastructure: $0.055/min
• TTS (Retell platform voices): $0.015/min
• LLM (GPT 5.1): $0.04/min
• Telephony (Retell Twilio): $0.015/min
• Knowledge base add-on: $0.005/min)

Monthly Retell AI cost: 2000 x ($0.055+$0.015+ $0.04+$0.015+$0.005) = $260/month

3. Outbound sales agent

Outbound campaigns add Batch Call and Branded Caller ID, plus a stronger LLM for objection handling and premium voices to improve answer rates.

• AI agent usage: 5,000 min / 2,000 dials per month
• Retell voice infrastructure: $0.055/min
• TTS (ElevenLabs voices): $0.04/min
• LLM (Claude 4.5 Sonnet): $0.08/min
• Telephony (Retell Twilio): $0.015/min
• Branded caller ID (2,000 calls): $0.10/call =$200
• Batch call (2000 dials): $0.005/dial = $10

Monthly Retell AI cost = 5000 x ($0.055+$0.04+$0.08+$0.015) +$200+ $10 = $1,160/month

These examples show why Retell AI pricing can vary significantly between businesses. Call volume, LLM selection, voice provider, telephony, and additional features all affect the final Retell AI billing amount.

Retell AI pricing vs. Other AI voice platforms

Retell AI competes with several AI voice platforms that use different pricing models. Here’s how Retell AI stacks up against five common alternatives.

Whether an alternative is better and cheaper than Retell AI depends on your call volume, model selection, telephony setup, and required features. 

• LiveKit

LiveKit is the strongest cheaper option, especially at scale. You can self-host its open-source server for free, eliminating per-minute platform fees entirely. Even on LiveKit Cloud, the $0.01/min agent-session fee undercuts Retell’s $0.055/min voice infrastructure charge.

• Vapi

Vapi runs close to Retell in practice, but its bring-your-own-key model can make it cheaper. If you already have provider accounts for STT, LLM, and TTS, Vapi’s platform fee starts at just $0.05/min.

• Synthflow

A practical option for businesses that want to build AI voice agents with less development work. Its visual workflows, integrations, and business-focused tools make it worth considering when ease of deployment and built-in business functions matter more than achieving the lowest per-minute rate.

• Bland AI

Bland AI is worth considering purely for budget predictability. Its published rates include the LLM, STT, and TTS components, which makes budgeting even easier. 

• Cognigy

Cognigy may not be cheaper for smaller deployments, but it can provide better value for organizations with complex contact-center requirements.

How to reduce your Retell AI costs

Because Retell AI pricing is modular, the components you choose affect your bill more than the base rate. A few practical changes can help keep Retell AI billing under control without affecting your call quality.

• Choose the right AI model for each use case

Use lower-cost models for simple tasks such as FAQs, appointment confirmations, or basic routing. Reserve premium models for conversations that require deeper reasoning.

• Keep prompts focused

Use concise system prompts and remove instructions, examples, or context the agent doesn’t need. Smaller prompts can reduce LLM usage and help control model-related costs.

• Reduce unnecessary agent turns

Design call flows so the agent collects related information in one go instead of repeatedly asking for small pieces of information. Fewer back-and-forth exchanges can reduce overall call duration.

• Set clear call termination rules

Define when the agent should end a call, such as after completing a booking, resolving a request, or detecting that the caller has finished. This helps prevent unnecessary call extensions.

• Monitor average call duration

Since billing is usage-based, a small increase in average call length across thousands of monthly calls compounds fast. Track this metric to find ways to reduce average call duration.

• Review usage regularly

Add-ons like AI Quality Assurance, Knowledge Base, or extra concurrency slots are easy to enable during testing and forget about. A monthly review can catch charges that no longer match actual usage.

• Compare telephony and voice-provider costs

Review the rates from your selected telephony and voice providers. Switching to a lower-cost option can reduce the overall Retell AI cost per minute when the quality and requirements remain suitable.

Retell AI pros and cons

Pros

• Transparent and publicly priced rates for each and every component
• Full platform access from day one without needing to upgrade to Enterprise
• Choose across multiple LLM and TTS providers based on your requirements and budget
• Build and validate a working agent for free with $10 in credits and 20 free concurrent calls
• Failed calls aren’t billed, and the AI agent fee stops the moment a call is handed off to a human.

Cons

• Final cost varies by your LLM configuration, voice provider, telephony, and any additional features you select
• No published Enterprise pricing, so budgeting requires a sales conversation
• Teams needing compliance features but not ready for full Enterprise scale have no option between pay-as-you-go and Enterprise
• Premium voice and model choices get expensive fast

Conclusion: Is Retell AI worth the price?

Retell AI is worth the price for teams that want control over their voice agent stack and are comfortable doing a bit of math. It isn’t the cheapest option on the market, but its pricing is transparent and predictable once you understand how the layers stack.

The real cost structure runs roughly $0.07-$0.31/min on Pay-as-you-go, built from four components: voice infrastructure, TTS, LLM, and telephony. Enterprise pricing is custom-quoted for teams needing dedicated infrastructure or compliance coverage. Model and voice choice, more than anything else, drive the final bill.

Consider a Retell AI alternative if you need a simpler bundled price, lower costs for a specific usage pattern, a visual agent-building experience, or features that better match your contact-centre requirements.