Disclosure: some links in this article are affiliate links. If you buy through them, Sophrai may earn a commission at no extra cost to you. It never changes our rankings. Read our affiliate disclosure.
Short answer: ElevenLabs Agents is the simpler and usually slightly cheaper bill: every paid plan works out to about $0.08 per call minute plus LLM tokens (from $0.0008/min), so 1,000 minutes cost about $80 + LLM on Creator, with 10 to 40 concurrent calls included by plan. Vapi charges $0.05/min hosting plus each provider at cost, which its calculator puts at $82 to $129 per 1,000 minutes, includes only 4 concurrent calls before you pay $10 per extra line, but lets you choose every provider, build multi-assistant squads and add HIPAA for $2,000/month. Carrier minutes (Twilio, Telnyx, SIP) are extra on both.
Prices from vapi.ai/pricing and the ElevenAgents tab of elevenlabs.io/pricing, checked 7 November 2026 (Asia/Tehran). We haven’t run paid production traffic on either platform.
Vapi and ElevenLabs Agents (ElevenLabs now brands it ElevenAgents) are two of the most common choices for B2B teams building AI phone agents for inbound support, appointment booking, lead qualification and outbound follow-up. They’re built differently. Vapi is an orchestration layer: you pick the speech-to-text, LLM, voice and telephony providers, and Vapi adds a per-minute hosting fee. ElevenLabs sells an all-in call minute that includes its own speech-to-text and text-to-speech, and bills the LLM on top. This comparison uses prices we checked on both official pricing pages on 7 November 2026, works out costs from 300 to 10,000 minutes a month, and shows how each connects to n8n. If you’re also weighing Retell AI or Bland, read our Vapi vs Retell AI vs Bland AI comparison.
Vapi vs ElevenLabs Agents at a glance
| Vapi | ElevenLabs Agents | |
|---|---|---|
| Pricing model | $0.05/min hosting + STT, LLM, voice at provider cost (no markup) | Plan with included call minutes, then $0.08/min; LLM billed separately |
| Typical cost per 1,000 min | $82–$129 (Vapi calculator, before carrier) | ~$80 on Creator + LLM ($0.80–$28.40) |
| Free start | $5 free credits, 1 free US number (inbound only) | Free plan: 15 call minutes/month |
| Entry paid plan | Core success package $29/month | Starter $6/month ($1 first month), 75 min |
| Included concurrency | 4 (usage only), 10 (Core), 30 (Pro) | 4 / 6 / 10 / 20 / 30 / 40 by plan |
| Extra concurrency | $10 per line per month | Burst up to 3× the limit (max 300) at $0.16/min |
| Telephony | Free Vapi SIP/WebSockets/WebRTC; Twilio $0.008 in / $0.014 out, Telnyx $0.0055 (billed by provider) | No ElevenLabs telephony fee; bring Twilio or a SIP trunk (billed by provider) |
| Bulk outbound | Outbound Campaigns (needs an imported number) | Batch calling from CSV/XLS (moving to Campaigns) |
| Multi-agent | Squads with handoff tool | Workflow builder (listed on Free and up) |
| Compliance | HIPAA add-on $2,000/month (DPA + BAA); ZDR from Core | BAAs for HIPAA on Enterprise |
| Raw data retention | 14 days (usage only), 30 (Core), 180 (Pro) | Not listed on the pricing page |
| n8n | Official n8n guide (API Request tool → Webhook) | Webhook tools + post-call webhooks |
Vapi pricing in detail
Vapi’s pricing page splits the bill into usage and an optional success package. Usage is Vapi’s $0.05 per minute hosting fee plus four components billed at the provider’s listed price with no markup: speech-to-text, the language model, text-to-speech and transport. The page’s default example shows Deepgram at $0.0095–$0.0099/min, OpenAI at $0.0077–$0.0452/min and an ElevenLabs voice at $0.0146–$0.0238/min, which the calculator totals at $82 to $129 a month for 1,000 minutes.
| Package | Price | Concurrent calls | Vapi numbers | Organizations | Raw data retention | Support |
|---|---|---|---|---|---|---|
| No package (usage only) | $0 + usage; $5 free credits | 4 | 1 | 2 | 14 days | Discord, email without SLA |
| Core | $29/month + usage | 10 | 5 | 2 | 30 days, ZDR included | Email, 2-business-day SLA |
| Pro | 10% of Vapi hosting fees, $999/month minimum | 30 | 10 | 10 | 180 days, RBAC + ZDR | Email + Slack, 1-business-day SLA, 99% uptime SLA |
| Premier | Custom (sales only) | Custom | Custom | Custom | Custom, SSO | 1-hour P0 SLA, 99.9% uptime SLA |
Add-ons work on any package: $10 per extra concurrent line per month, $20 per extra organization per month, and HIPAA-eligible data handling for $2,000 per month per organization, which comes with a standard DPA and BAA. Transport is free on Vapi Telephony/SIP, Vapi WebSockets and Daily WebRTC; carriers bill their own rates, which Vapi lists as Twilio $0.008/min inbound and $0.014/min outbound, Vonage $0.00814 and Telnyx $0.0055.
Two limits matter for sales teams. According to Vapi’s free-number docs, the free Vapi number uses a US area code, takes inbound calls only and can’t be ported out, so outbound calls and Outbound Campaigns need a number imported from Twilio, Telnyx or another provider. And concurrency is measured per subscription and shared by all its organizations.
ElevenLabs Agents pricing in detail
On the ElevenAgents tab of ElevenLabs’ pricing page, agent plans are billed by call minutes, not by the shared credit pool used for other ElevenLabs products:
| Plan | Monthly price | Included call minutes | Concurrent calls | Effective $/min (our arithmetic) |
|---|---|---|---|---|
| Free | $0 | 15 | 4 | — |
| Starter | $6 ($1 first month) | 75 | 6 | $0.080 |
| Creator | $22 ($11 first month) | 275 | 10 | $0.080 |
| Pro | $99 | 1,238 | 20 | $0.080 |
| Scale | $299 (3 workspace seats) | 3,738 | 30 | $0.080 |
| Business | $990 (10 seats) | 12,375 | 40 | $0.080 |
| Enterprise | Custom | Custom | Elevated | Custom |
Additional call minutes cost $0.08 per minute, text messages $0.003 each, and burst calls above your concurrency limit $0.16 per minute (up to 3× your limit, maximum 300 calls for non-enterprise customers). The interesting part is the last column: every paid plan’s included minutes are priced at almost exactly the overage rate, so higher plans don’t buy cheaper minutes. They buy concurrency, seats and voice-clone slots.
LLM usage comes on top and is deducted from your ElevenLabs credits. The cost calculator on the same page showed these per-minute estimates on 7 November 2026: Gemini 2.5 Flash Lite $0.0008, GPT-4o Mini $0.0012, GPT-5 Mini $0.0024, Gemini 2.5 Flash $0.0029, GPT-4.1 Mini $0.0033, Claude Haiku 4.5 $0.0095 and Claude Sonnet 4.6 $0.0284. ElevenLabs adds no telephony fee: phone calls cost the same call-minute rate as any other channel, and your Twilio or SIP provider bills the carrier side directly.
What does each cost per month?
Platform, speech and voice cost before carrier charges. Vapi uses its calculator’s per-minute range ($0.082–$0.129) scaled linearly; ElevenLabs uses the cheapest plan that covers the minutes, with GPT-4.1 Mini as the LLM.
| Minutes per month | Vapi (usage only) | ElevenLabs Agents (cheapest plan) | ElevenLabs LLM (GPT-4.1 Mini) | ElevenLabs total |
|---|---|---|---|---|
| 300 | $24.60–$38.70 | $24.00 (Creator $22 + 25 × $0.08) | $0.99 | ~$25 |
| 1,000 | $82–$129 | $80.00 (Creator $22 + 725 × $0.08) | $3.30 | ~$83 |
| 5,000 | $410–$645 | ~$400 (Scale $299 + 1,262 × $0.08) | $16.50 | ~$416 |
| 10,000 | $820–$1,290 | ~$800 (Scale $299 + 6,262 × $0.08) | $33.00 | ~$833 |
Our arithmetic from official list prices on 7 November 2026. Excludes Twilio/Telnyx/SIP carrier minutes and phone numbers, taxes, Vapi success packages and ElevenLabs’ first-month discounts. A premium LLM such as Claude Sonnet 4.6 adds about $28 per 1,000 minutes on ElevenLabs; on Vapi, the LLM is already inside the $82–$129 range.
Three takeaways. First, at low and mid volume the two land within a few dollars of each other if you run Vapi with cheaper models; the gap opens when Vapi uses premium LLMs and voices. Second, concurrency is where ElevenLabs is clearly cheaper: 20 simultaneous calls come with Pro at $99, while on Vapi you’d pay Core $29 plus 10 extra lines at $10, or $129 a month before usage. Third, Vapi’s cost is more controllable: you can swap in a cheaper speech-to-text or voice provider, which an all-in minute doesn’t allow.
Features that matter for B2B voice agents
Provider choice and voice quality
Vapi’s catalogue lets you mix providers per assistant, and its pricing page lists ElevenLabs among the voice options, so you can run ElevenLabs voices on Vapi. ElevenLabs Agents uses its own speech models and a choice of hosted LLMs or a custom LLM; its agents overview describes 10,000+ voices. If voice quality is the deciding factor, test the same script on both, because both can end up using ElevenLabs voices.
Outbound calling and campaigns
For outbound sales calls, Vapi needs an imported number for both single outbound calls and Campaigns. ElevenLabs’ batch calling uploads CSV or XLS lists with dynamic variables and schedules them, works with the native Twilio integration and SIP trunking, and uses the lower of 50% of workspace concurrency or 70% of the agent’s limit so inbound calls still get through. ElevenLabs flags batch calling for deprecation in favour of Campaigns, and Zero Retention Mode can’t be used with batch calls.
Tools, MCP and webhooks
Both platforms let agents call your APIs during a conversation and both support MCP. ElevenLabs’ webhook tools support custom headers and auth connections, and its MCP integration connects external tool servers. After the call, ElevenLabs sends post-call webhooks of three types (transcription, audio and call-initiation failure) signed with an HMAC ElevenLabs-Signature header. Vapi’s API Request tool calls any HTTPS endpoint, and its Squads hand a call between specialised assistants.
How to connect each to n8n
Vapi publishes an official n8n guide: create an API Request tool pointing at the production URL of an n8n Webhook node, set the Webhook to respond with a Respond to Webhook node returning JSON, and keep the workflow faster than the tool timeout (timeoutSeconds: 10 in Vapi’s starter). Use the Test URL only while n8n is listening.
For ElevenLabs, add a webhook tool that calls your n8n Webhook URL during the call for lookups (CRM record, calendar slot, order status), then point post-call webhooks at a second n8n workflow that verifies the HMAC signature, writes the transcript and outcome to your CRM and triggers follow-up email or Slack alerts. Use the HTTP Request node to start calls or batches through either API. Self-hosting n8n? Size the server with our n8n cost calculator and check finished workflows with the n8n workflow auditor.
Which should you choose?
- Choose ElevenLabs Agents if you want one predictable rate (about $0.08/min plus LLM), need 10–40 concurrent calls without add-ons, or are a small team that wants to launch inbound support or booking quickly on Twilio.
- Choose Vapi if you’re a developer or agency that wants to choose and swap every provider, needs Squads for multi-step call flows, wants a HIPAA BAA without an enterprise contract, or plans to serve several clients from separate organizations.
- Test both if call quality decides the deal: run the same 50 test calls with the same LLM and voice, and compare latency, interruptions and tool-call success.
Comparing more AI tools? Our AI SaaS pricing comparison puts list prices side by side, and our Cartesia Managed Agents review covers a third option at $0.03 per minute.
What we couldn’t verify
- ElevenLabs annual prices for agent plans. The yearly toggle didn’t change the ElevenAgents table for us. The general plan FAQ says annual billing equals 10 months’ price (for example Pro $82.50 a month), but we didn’t confirm that it applies identically to agent minutes.
- Overage on Free and Starter. “Additional Minutes” is listed as a feature from Creator upward, so Free and Starter appear to stop at their included minutes. The pricing table shows $0.08 for every plan; check in the dashboard.
- Exact Vapi cost for your stack. Vapi’s provider prices depend on the models and voices you pick; we used its calculator’s default range and scaled it linearly.
- Language counts. ElevenLabs’ agents overview mentions both 70+ languages and “31 languages” in different places. Test the languages you need.
- Latency and call quality. We didn’t benchmark either platform; any latency figures on their sites are vendor claims.
- Vapi Core sign-up. The Core card shows $29/month pre-paid but also a Contact Sales button; self-serve availability may vary by account.
Frequently asked questions
Is Vapi or ElevenLabs Agents cheaper?
For most volumes ElevenLabs Agents is slightly cheaper on platform and voice cost. Every paid ElevenAgents plan works out to about $0.08 per call minute (included minutes and overage), plus LLM usage, which ElevenLabs’ calculator shows at $0.0008 to $0.0284 a minute depending on the model. Vapi charges $0.05 a minute hosting plus speech-to-text, LLM and voice at provider cost; its own calculator estimates $82 to $129 per 1,000 minutes. At 1,000 minutes that’s about $80 plus LLM on ElevenLabs Creator versus $82 to $129 on Vapi. Carrier charges (Twilio, Telnyx, SIP) are extra on both.
How many concurrent calls do Vapi and ElevenLabs allow?
Vapi includes 4 concurrent calls on usage-only billing, 10 on the $29 Core package and 30 on Pro ($999 a month minimum); extra lines cost $10 per line per month. ElevenLabs includes 4 on Free, 6 on Starter, 10 on Creator, 20 on Pro, 30 on Scale and 40 on Business, and burst pricing lets agents go up to 3 times the limit (maximum 300 calls for non-enterprise customers) at $0.16 a minute.
Do Vapi and ElevenLabs charge for phone numbers and telephony?
Vapi gives one free US Vapi number (inbound only; Core includes 5, Pro 10) and lists carrier rates such as Twilio $0.008 a minute inbound and $0.014 outbound, which the provider bills, not Vapi. ElevenLabs says telephony is included on every plan with no extra ElevenLabs fee; you connect Twilio or a SIP trunk and that provider bills you directly.
Which is better for outbound calling campaigns?
Both support bulk outbound calls. Vapi has Outbound Campaigns, but they need an imported number from a provider such as Twilio or Telnyx, because free Vapi numbers are inbound only. ElevenLabs has batch calling from CSV or XLS lists, which uses the lower of 50% of your workspace concurrency or 70% of the agent’s limit, and it says batch calling will be replaced by Campaigns. Check consent and TCPA rules before any outbound campaign.
Can I connect Vapi or ElevenLabs Agents to n8n?
Yes. Vapi’s docs include an n8n guide: an API Request tool calls an n8n production Webhook URL, and a Respond to Webhook node returns JSON, with a 10-second tool timeout in the starter. ElevenLabs agents call n8n through webhook tools during the call and can send post-call webhooks (transcription, audio and call-initiation-failure events, signed with an HMAC ElevenLabs-Signature header) to an n8n Webhook node after it.
Which should I choose?
Choose ElevenLabs Agents if you want the simplest bill (about $0.08 a minute plus LLM), more concurrency per dollar and ElevenLabs voices built in. Choose Vapi if you want to pick and swap every provider (speech-to-text, LLM, voice, telephony), need multi-assistant squads or a HIPAA add-on without an enterprise contract, or want a free inbound US number to start.
Related reading
- Vapi vs Cartesia Managed Agents: voice pricing compared
- ElevenLabs vs Cartesia: TTS and voice agent pricing compared
- Vapi vs Retell AI vs Bland AI: voice agent pricing and limits
- Cartesia Managed Agents at $0.03 per minute
- WhatsApp Business API pricing changes (October 2026)
- Attio vs Close: CRM pricing for B2B sales
- Lindy vs Relevance AI vs n8n for B2B AI agents
Sources and method
Prices and limits come from Vapi pricing (including its FAQ) and the ElevenAgents tab of ElevenLabs pricing (plan table, cost calculator and FAQ), both checked 7 November 2026 (Asia/Tehran). Feature details come from Vapi’s free telephony, call concurrency, Outbound Campaigns and n8n docs, and ElevenLabs’ agents overview, webhook tools, MCP, post-call webhooks and batch calling docs, plus n8n’s Webhook and HTTP Request node docs. Monthly totals are our own arithmetic. Prices change, so check both pages before you commit.
