Bland vs Retell vs Vapi
Developer voice-agent stacks: all three advertise a floor, not a bill. Scores from the RXed AI Periodic Table v1.0 — identical 20-element scale, deterministic overall. How we audit →
At a glance
Scope = elements of the AI stack in the tool’s stated scope (dashed = honestly out of scope, never a penalty). Quality = how well it scores on what it covers. No single number — both axes, always.
The verdict
Bland (Scope 17/20 · Quality 7/10) owns its entire stack — speech-to-text, inference, text-to-speech — and has the testing rigour to match: evals with LLM judges (Ev 9.5) and node-level regression tests, at the cost of the trio's highest measured latency (~700-900ms). Retell (Scope 16/20 · Quality 7/10) is the code-level control pick: real MCP server, real fine-tuning, a live pricing calculator on the homepage — just remember $0.07/min is the floor and production runs $0.11-0.31/min. Vapi (Scope 18/20 · Quality 6/10) has the widest scope and real infrastructure behind the Amazon Ring deployment, but budget 3-7x the advertised $0.05/min once STT, LLM, TTS and telephony all send separate invoices. Regulated, high-stakes phone work → Bland. Fast iteration with model choice → Retell. Maximum stack control, multi-vendor ops tolerated → Vapi.
Where the quality sits
Family mean over covered elements only — a dash means the family is honestly outside that tool’s scope.
| Bland Scope 17/20 · Quality 7/10 audited 2026-08-06 | Retell AI Scope 16/20 · Quality 7/10 audited 2026-08-05 | Vapi Scope 18/20 · Quality 6/10 audited 2026-07-25 | |
|---|---|---|---|
| Vendor | Bland (Bland AI, Inc.) | Retell AI | Vapi, Inc. |
| Origin | US | US | US |
| Pricing | Start free (~€0.12/min talk time) · Build ~€257/month (~€0.10/min) · Scale ~€429/month (~€0.09/min) · Enterprise contracted to volume — converted from official USD pricing at ~0.86 EUR/USD, checked 06/08/2026. Subscriptions include zero minutes; usage is billed on top. | $0.07-$0.31/min voice (~€0.065-€0.29/min), true pay-as-you-go with $10 free credits and no platform fee; real blended production cost typically €0.10-€0.29/min once LLM and telephony are stacked on. Chat agents from $0.002+/message. Enterprise: custom. | Build (pay-as-you-go): ~€0.044/min (~$0.05) hosting + model-provider costs at cost (€0 with your own API key); 10 concurrent calls included, +~€8.75/line/mo (~$10); HIPAA add-on ~€1,750/mo (~$2,000); Zero Data Retention add-on ~€875/mo (~$1,000) · Scale: annual contract, custom volume-based pricing (SOC2, HIPAA, PCI, SSO, RBAC, dedicated support). Verified on vapi.ai/pricing 25/07/2026; USD-denominated, converted at 1 USD = 0.875 EUR. |
| Pr Prompts | 9 | 9 | 8 |
| Em Embeddings | — | — | 4 |
| Cx Context | 8 | 7 | 7 |
| Tr Tracing | 9 | 8 | 7 |
| Lg LLM | 7 | 8 | 8 |
| Fc Function calling | 8 | 8 | 9 |
| Vx Vector store | 6 | — | 3 |
| Rg RAG | 8 | 8 | 6 |
| Gr Guardrails | 9 | 8 | 6 |
| Mm Multimodal | 9 | 6 | 5 |
| Ag Agents | 8 | 8 | 9 |
| Ft Fine-tuning | 7 | 7 | 4 |
| Fw Frameworks & harnesses | 9 | 8 | 8 |
| Ev Evaluations | 10 | 8 | 7 |
| Sm Small models | — | 8 | 6 |
| Ma Multi-agent | 6 | 4 | 8 |
| Sy Synthetic data | 5 | — | — |
| Pc Protocols | 8 | 7 | 8 |
| In Interpretability | 3 | — | — |
| Th Thinking models | — | 6 | 5 |
Hover any score for the written reason. Full detail: Bland passport → · Retell AI passport → · Vapi passport →