✉ The Friday AI Brief: the week's 5 best AI stories, tools & comparisons — in your inbox every Friday morning.

Bland AI logo

Bland AI Review 2026: One Flat Per-Minute Rate for AI Phone Agents — With a Hidden Fourth Plan

Research reviewed November 2026. This is a research-based review — we analyzed published specs, pricing, and user feedback.

Quick verdict

✓ Best for:High-volume, regulated phone operations that want AI agents on one flat per-minute rate with the model costs included.
✕ Skip it if:You run only a handful of calls a day: Build adds a $299/mo platform fee, and the cheapest $0.11/min rate is on a docs-only Scale plan.

TL;DR: Bland AI is an enterprise voice AI platform that builds and runs AI phone agents on its own speech stack — its own transcription, language model, and voices — and charges one flat per-minute rate with the model costs included. Start is genuinely free to try at $0.14/min with no platform fee, but serious volume runs into a $299/month platform fee on Build, and the cheapest per-minute rate ($0.11/min) sits on a Scale plan you will not find on the public pricing page. It is a strong fit for high-volume, regulated phone work; it is overkill for a small business that just wants a receptionist.

What Bland AI does

Bland builds AI agents that make and take real phone calls: inbound support and scheduling, outbound lead qualification, and reminders. Where rivals such as Retell AI and Vapi assemble third-party models and telephony, Bland runs the whole stack itself — speech-to-text, the language model, text-to-speech, and call orchestration — its main pitch to regulated industries like healthcare, insurance, and financial services. Agents are built from a prompt, optional Conversational Pathways for branching call flows, tools, and knowledge bases. The company lists SOC 2 Type II, HIPAA, GDPR, and PCI DSS coverage, with on-premise and VPC deployment on Enterprise.

Bland AI pricing

Plan Platform fee Talk rate Transfer rate Limits
Start $0 $0.14/min $0.05/min 10 concurrent calls, 100 calls/day, 10 knowledge bases
Build $299/mo $0.12/min $0.04/min 50 concurrent calls, 2,000 calls/day, 50 knowledge bases
Scale (docs only) $499/mo $0.11/min $0.03/min 100 concurrent calls, 5,000 calls/day, 15 voice clones
Enterprise Custom Custom Custom Unlimited; BAA, SSO, data residency, on-prem/VPC

The per-minute rate includes the language model (no token charges), real-time transcription, and premium voices. Telephony is billed separately — your own carrier, or Bland’s at pass-through cost — and if you bring your own telephony, transfer fees drop away entirely. Start includes 2 credits plus an inbound number Bland values at $15/month, with no card required.

What the pricing page doesn’t make obvious

Bland’s public pricing page shows only Start, Build, and Enterprise. Its billing documentation lists two more tiers: Scale ($499/month at $0.11/min) and an Agent Phone plan at $29.99 for a single agent number. The docs also price the small print: outbound attempts and failed calls on Bland’s telephony carry a $0.015 minimum per attempt, SMS costs $0.02 per message in both directions, and web-widget messages cost $0.01 each. The current rates are also a recent increase — until December 5, 2025, every plan paid a flat $0.09/min.

Strengths

  • One predictable number: the per-minute rate covers the model, transcription, and voice, so there are no stacked provider bills to forecast.
  • A real free on-ramp: Start costs nothing to join, needs no card, and includes an inbound number.
  • Compliance posture: its own infrastructure means one vendor covers the whole call, which matters for HIPAA and similar reviews.
  • Scale headroom: published caps run to 5,000 calls a day before you are pushed to Enterprise.

Weaknesses

  • The platform-fee step is steep: going from free Start to Build is $299/month before a single minute is used.
  • Locked voices and models: because the stack is Bland’s own, you cannot swap in a rival’s voice or model the way modular platforms allow.
  • Extras still add up: telephony, transfers, SMS, and per-attempt minimums sit outside the headline rate.
  • Enterprise gating: BAA, SSO, and data residency sit on the custom tier, and compliance documents are shared under NDA on Enterprise — unlike Synthflow’s enterprise-only contracts, you can start self-serve, but the serious compliance conversation still ends with sales.

Who Bland AI is for

Operations, healthcare, insurance, and lending teams running hundreds or thousands of calls a day, who want one vendor and one rate card. Developers can prototype free on Start. Small businesses with a handful of daily calls should look elsewhere first — the Build platform fee only pays for itself at real volume.

Verdict

Bland AI’s flat, all-in per-minute pricing is the clearest rate card in voice AI, and its owned stack is a genuine differentiator for regulated work. Just budget from the billing docs, not the pricing page: the plan you probably want at scale, and several small per-call charges, only appear there.

FAQ

How much does Bland AI cost per minute?

$0.14/min on the free Start plan, $0.12/min on Build ($299/month platform fee), and $0.11/min on the Scale plan ($499/month) listed in Bland’s billing docs. Telephony, transfers, and SMS are extra.

Is Bland AI free to try?

Yes. Start has no platform fee and no card requirement, and includes 2 credits plus an inbound number.

Does Bland AI charge for failed calls?

On Bland’s own telephony, yes — outbound attempts and failed calls carry a $0.015 minimum charge per attempt, per its billing documentation.

How is Bland different from Retell AI or Vapi?

Bland runs its own transcription, model, and voices and bundles them into one per-minute rate. Retell and Vapi charge a lower platform rate but bill the underlying model, voice, and telephony providers separately.

Leave a Comment

Your email address will not be published. Required fields are marked *

Get the 5 best AI tools every week

Top AI news, tools, and prompts — one short email. Free, unsubscribe anytime.

Run a newsletter of your own? Monetize and grow it with SparkLoop →

Scroll to Top