# Bland Speech vs Google instant custom voice cloning (2026): What is still open to a new team

Google says a team can no longer request access to Chirp 3: Instant Custom Voice, and it points developers to voice replication with Gemini 3.8 TTS instead. That sends a team from a product with a price list and a consent flow to one labelled Preview, which Google says is available through its Agent Platform API and Agent Studio and not through the Cloud Text-to-Speech API.

A buyer comparing Google with Bland Speech for a cloned voice is therefore comparing a closed product and a preview with a product that sells clones today. Bland's [text to speech comparison hub](/compare/text-to-speech) and [voice cloning comparisons](/compare/text-to-speech/voice-cloning) cover the other vendors.

## Which is better, Bland Speech or Google, for voice cloning?

For a team that needs to start a cloned voice now, Bland Speech is the better fit as of November 2026: Google's instant custom voice is closed to new requests, Gemini 3.8 voice replication is in Preview, and Bland Speech v3 ranked first of 13 AI voices in Design Arena's Audio Realism Bench results of 4 August 2026. Google is the better fit for a team already inside Google Cloud that will build on Gemini 3.8 voice replication while it is in Preview.

<div class="cmp-usewhen"><div class="cmp-usewhen-own"><p><strong>Use Bland Speech when</strong></p><ul><li>A team will start a new clone from one clean sample of about 10 seconds, or from 30 minutes or more of audio for a professional clone</li><li>A team will confirm that every cloned voice is its to use and keep each clone private to its organization</li><li>Cloned speech will go straight onto a phone line as mu-law at 8 kHz</li><li>A voice agent will stream text from a language model into a WebSocket and let callers interrupt mid-sentence</li></ul></div><div><p><strong>Use Google when</strong></p><ul><li>A team will build on a Preview feature, Gemini 3.8 voice replication, through Google's Agent Platform API</li><li>A team will accept Preview terms in return for Google's token pricing</li><li>A team will run Text to Speech under Google Cloud's HIPAA business associate agreement</li><li>A script will carry tags such as [sigh], [laughing], and [short pause] with a style prompt for tone</li></ul></div></div>

| Feature | Bland Speech | Google |
|---|---|---|
| Audio Realism Bench, 4 August 2026 | Ranked 1st of 13 AI voices, 1365 | Gemini 2.5 Pro TTS ranked 5th, 1109 |
| Cloning open to a new team | Yes | Chirp 3 Instant Custom Voice closed, Gemini 3.8 voice replication in Preview |
| Audio for an instant clone | One clean sample, about 10 seconds | 10 seconds |
| Consent check | Customer confirms each voice is its to use, plus a consent and rights review for professional clones | A required consent statement recorded in the chosen language |
| Price per million characters | $40, one standard rate | $60 for instant custom voice, with no free usage |
| Free usage | 50,000 characters, once | None on instant custom voice |
| Phone audio | Mu-law at 8 kHz | Mu-law and A-law |

<details><summary>What these rows mean</summary>

- **Audio Realism Bench, 4 August 2026:** Design Arena's blind listening test, where listeners hear two clips of the same line from two models and pick the one that sounds more like a real person. Google's models in the results were Gemini 2.5 Pro TTS (5th, 1109), Gemini 3.1 Flash TTS (8th, 1037), and Gemini 2.5 Flash TTS (10th, 985). Chirp 3 and Gemini 3.8 weren't in them. Each model was scored on its stock voices, so the numbers describe a model, not any one clone.
- **Cloning open to a new team:** Google says "You can no longer request access to Chirp 3: Instant Custom Voice. Instead, use voice replication with Gemini 3.8 TTS." Google marks Gemini 3.8 voice replication as Preview.
- **Audio for an instant clone:** Bland describes one clean sample of about 10 seconds. Google says instant custom voice needs just 10 seconds of audio.
- **Consent check:** Google requires a recorded consent statement in the chosen language, with no custom script allowed. Bland asks the customer to confirm every voice is theirs to use, keeps every clone private to the customer's organization, and reviews professional clones for consent and rights.
- **Price per million characters:** Bland Speech costs $0.04 per 1,000 characters, which is $40 per million. Google charges $0.00006 per character for instant custom voice, which is $60 per million, and states no free usage.
- **Phone audio:** Bland returns mu-law at 8 kHz. Google's Cloud Text-to-Speech returns MULAW and ALAW encodings, and Gemini-TTS streaming offers MULAW and ALAW too.

</details>

## How do Bland Speech and Google voice cloning compare?

Bland Speech and Google differ most on what a new team can start. Bland Speech sells instant and professional clones today, while Google's route to a cloned voice has changed under it and now ends in a Preview.

### What is open to a new team today?

Google's Chirp 3 Instant Custom Voice no longer takes access requests, and Google points to voice replication with Gemini 3.8 TTS. Google says Gemini 3.8 Flash TTS and Flash-Lite TTS are generally available through its Agent Platform API and Agent Studio, with voice design and voice replication in Preview, and that they aren't available through the Cloud Text-to-Speech API.

Bland Speech makes an instant clone from one clean sample of about 10 seconds and a professional clone from 30 minutes or more of audio.

### How does each handle consent?

Google requires a consent statement for instant custom voice, recorded in the language the voice will speak, and it won't accept a custom script in place of the default. That is a fixed check. Bland has the customer confirm that every voice is theirs to use, keeps each clone private to the customer's organization, and puts professional clones through a consent and rights review before training.

### Bland Speech cloning overview

Bland Speech is the text to speech model inside Bland, a voice AI provider that builds and runs its own speech model for enterprise phone calls. Bland's default cloning engine is BTTS V3, sold as part of [Bland Speech](/speech). Bland's [text to speech API](/text-to-speech-api) takes POST /v2/tts for complete text and WSS /v2/tts/ws for streamed text.

<div class="cmp-proscons cmp-proscons-own"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>An instant clone from one clean sample of about 10 seconds</li><li>A professional clone fine-tuned on 30 minutes or more of audio, after a consent and rights review</li><li>Every clone private to the customer's organization</li><li>No monthly plan, with $0.04 per 1,000 characters pay as you go</li><li>Mu-law at 8 kHz and PCM from 8 to 48 kHz for phone lines</li><li>Bland Speech v3 ranked first of 13 AI voices on Audio Realism Bench on 4 August 2026</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>More than a hobby project needs, because Bland Speech is built for production calls</li></ul></div></div>

### Google voice cloning overview

Google Cloud sells Text-to-Speech with Chirp 3: HD voices, Gemini-TTS models, and legacy WaveNet, Neural2, and Studio voices, across 75+ languages and variants. Text-to-Speech is covered by Google Cloud's HIPAA business associate agreement and sits in its SOC 2 scope. Instant custom voice, which Google says needs 10 seconds of audio, is closed to new requests.

<div class="cmp-proscons" data-rival="google"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>A mandatory recorded consent statement for instant custom voice</li><li>Text-to-Speech covered by Google Cloud's HIPAA business associate agreement</li><li>Gemini-TTS inline tags and a style prompt for tone</li><li>Mu-law and A-law output</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Chirp 3 Instant Custom Voice is closed to new access requests</li><li>Gemini 3.8 voice replication is in Preview, through the Agent Platform API and not the Cloud Text-to-Speech API</li><li>Instant custom voice costs $60 per million characters with no free usage</li><li>Gemini 2.5 Pro TTS ranked 5th of 13 AI voices on Audio Realism Bench on 4 August 2026</li></ul></div></div>

## How much does a cloned voice cost on Bland Speech and Google?

Bland Speech costs $40 per million characters, and Google's instant custom voice costs $60, though a team starting today can't buy the Google product. Gemini 3.8 Flash TTS, the model family Google points to for replication, is priced in tokens, which comes to less per hour of audio than Bland Speech.

### Bland Speech pricing

Bland Speech costs $0.04 per 1,000 characters in the studio or the API, with no monthly plan. New accounts get 50,000 characters free, and credits don't expire. Enterprise pricing is custom, set with Bland's sales team.

### Google pricing

Google charges $0.00006 per character for instant custom voice, which is $60 per million, and Chirp 3: HD voices cost $30 per million with the first million characters free each month. Gemini 3.8 Flash TTS in Preview costs $0.50 per million text tokens in and $9.00 per million audio tokens out through 31 December 2026, then $1.00 and $18.00 from 1 January 2027. Google counts 25 audio tokens per second of audio, so output works out to about $0.81 for an hour of speech until the end of 2026, before input tokens.

Bland says 50,000 characters is nearly an hour of speech, which puts Bland Speech at about $2 an hour at $40 per million characters. Google's token price for Gemini 3.8 Flash TTS is lower per hour, and Google marks voice replication as Preview.

### What a month of instant custom voice costs

These totals use Google's published instant custom voice rate, which a new team can no longer request.

- **100,000 characters a month:** Bland Speech $4, Google $6.
- **1 million characters a month:** Bland Speech $40, Google $60.
- **10 million characters a month:** Bland Speech $400, Google $600.

## Which sounds more human, Bland Speech or Google?

Bland Speech v3 scored 1365 and ranked first of 13 AI voices in the blind tests Design Arena published on 4 August 2026, 256 points above Google's best result, Gemini 2.5 Pro TTS, which ranked 5th at 1109.

Listeners hear two clips of the same line without knowing which model made which and pick the one that sounds more like a real person. Chirp 3 and Gemini 3.8, the models behind Google's cloning, weren't in those results. Design Arena scored one female and one male stock voice per model, so [Design Arena's published results](https://notes.designarena.ai/audio-realism-benchmark-measuring-realism-in-audio-models/) describe how a model sounds, and a particular clone can land above or below it.

## Why is Bland Speech the best Google voice cloning alternative for phone calls?

A team picks Bland Speech over Google for a cloned phone voice when it wants a product that is open to new customers rather than a closed one and a preview, a voice that ranked first among AI voices on Audio Realism Bench on 4 August 2026, and one standard per-character rate. Both return mu-law, so the phone format isn't the difference.

A regulated team can weigh one more fact. Bland is SOC 2 Type II audited, HIPAA compliant with a signed BAA, and FedRAMP 20x Class C certified at Moderate, granted on 17 November 2026, and Bland says speech synthesis runs inside the FedRAMP boundary.

A voice cloned on Google doesn't move across. A team clones it again on Bland Speech from the original recordings with the voice owner's consent, and the [guide to moving a cloned voice to Bland Speech](/blog/move-cloned-voice-to-bland-speech) covers the steps. A team weighing Google's whole speech service can read [Google text to speech alternatives](/blog/google-text-to-speech-alternatives) or [how to migrate from Google text to speech](/blog/migrate-from-google-tts).

## Frequently asked questions

### Can a new team still get Google's instant custom voice?

No. Google says that a team can no longer request access to Chirp 3: Instant Custom Voice, and points to voice replication with Gemini 3.8 TTS, which Google marks as Preview.

### Is Bland Speech better than Google for voice cloning?

For a new cloned voice on phone calls, yes. Google's instant custom voice is closed to new requests, and Bland Speech v3 ranked first of 13 AI voices on Audio Realism Bench on 4 August 2026 where Gemini 2.5 Pro TTS ranked 5th.

### How much audio does Google need to clone a voice?

Google says instant custom voice takes 10 seconds of audio, along with a recorded consent statement in the voice's language. Gemini 3.8 voice replication is in Preview.

### How much does Google's instant custom voice cost?

Google charges $0.00006 per character, which is $60 per million characters, and states no free usage for it.

### How much audio does Bland Speech need to clone a voice?

An instant clone needs one clean sample of about 10 seconds. A professional clone is fine-tuned on 30 minutes or more of audio and passes Bland's consent and rights review.

### Does Google's text to speech work on phone calls?

Yes. Google's Cloud Text-to-Speech returns MULAW and ALAW encodings. Bland Speech returns mu-law at 8 kHz or PCM from 8 to 48 kHz, and Bland built it for phone calls.

### Can a Google clone move to Bland Speech?

No file moves across. A team clones the voice again on Bland Speech from the original recordings and confirms it is the team's to use.
