Azure AI Speech vs Google Cloud Text-to-Speech (2026): which should you pick?
As of Nov 12, 2026, Google Cloud Text-to-Speech grades higher than Azure AI Speech: 2.8 against 2.2 out of 4.0.
Checked Nov 12, 2026 · graded Nov 12, 2026 by ToolGPA
Google Cloud Text-to-Speech grades higher on value for money and features. They get the same grade on the free plan, ease of use and platforms.
Price, platforms and features side by side. The table says where each price was read and when.
Azure AI Speech
Microsoft speech API for text to speech, transcription and translation
Free plan
Google Cloud Text-to-Speech
Pay-per-character speech API with Gemini, Chirp and WaveNet voices
Free plan
Which one is for you
Pick Azure AI Speech if
- Fits if you already build on Azure and want speech in the same bill
- Works for developers who need transcription, translation and speech output from one vendor
Look elsewhere if
- Not for you if you want an editor to paste a script into, because it's an API and console service
- Tight if you plan custom voices, since access is limited and training and endpoint hosting cost extra
- Watch out: prices follow the region you pick and are labelled estimates that vary by agreement and currency
- Watch out: the product was renamed to Azure Speech in Foundry Tools, so older docs and tutorials use the earlier name
Pick Google Cloud Text-to-Speech if
- Fits if you are on Google Cloud and need speech output in an app or voice agent
- Works for long narration where a per-character price matters
Look elsewhere if
- Not for you if you want an editor to type into, because there is none
- Skip it if you can't add billing, since nothing works until it is enabled
- Watch out: instant custom voice has no free allowance and costs $60 per 1M characters
- Watch out: Gemini 3.8 models are Preview, with rates that rise on January 1, 2027
- Pricey if you pick Studio voices at $160 per 1M characters
How we grade them
| 2.2GPAAzure AI SpeechFair | 2.8GPAGoogle Cloud Text-to-SpeechGood | |
|---|---|---|
| Value for money | CNo flat plan. Neural voices run about $15 per 1M characters and Neural HD about $22 in US regions, so what you pay depends on volume and region. The free allowance is the only thing that brings it down. | AStandard and WaveNet are $4 per 1M characters, Neural2 $16, Chirp 3 HD $30 and Studio $160. With 4M characters free every month on the cheap voices, most small projects pay nothing, which is why this is graded against its category on the allowance. |
| Free plan and trial | BThe F0 tier gives 0.5M Neural characters a month plus 5 hours of speech to text, free. It sits behind an Azure account, so you set that up first. | B4M free characters on Standard and WaveNet and 1M on Neural2, Studio, Polyglot and Chirp 3 HD, every month, plus up to $300 credit for new customers. Billing must be enabled first, so it is not card-free. |
| Features | CNeural and Neural HD voices, custom and personal voice, transcription, translation, avatars and a real-time voice API. That's the job and then some, but custom voice is limited-access and training and hosting are billed separately. | BVoices across several tiers, Chirp and Gemini models, instant custom voice at $60 per 1M characters. It covers the job of speech output, though custom voice has no free allowance and Gemini rates are per token, with Preview rates rising January 1, 2027. |
| Ease of use | DNo store rating on record, and it's an API and console service. You write code or wire it into other Azure pieces, so this is not a start-typing tool. | DNo store rating on record. You make a Cloud project, turn on billing and call the API or use the console. That is developer setup. |
| Platforms and integrations | BWeb and API only, with Microsoft Foundry and Azure OpenAI as the listed hooks. No desktop or phone app. | BWeb and API, with Google Cloud and Gemini listed. No desktop or phone app. |
| Best for | speech features built on Azure | a pay-per-character speech API at volume |
The GPA is the average of the five grades: an A is 4 points, an F is none. The full reasons are on each tool's page: Azure AI Speech and Google Cloud Text-to-Speech. How we grade
Where they differ
- Azure AI Speech also covers transcription.
In commonFree planWebPublic APIText to speechVoice cloning
At a glance
| Azure AI Speech | Google Cloud Text-to-Speech | |
|---|---|---|
| Pricing | Free plan | Free plan |
| Starts at | Not published | Not published |
| Free plan | Free (F0) tier: 0.5 million Neural text-to-speech characters per month, plus 5 hours of standard speech to text. | Free monthly characters: 4 million on Standard and WaveNet, 1 million on Neural2, Studio, Polyglot and Chirp 3 HD voices. |
| Free trial | None listed | New Google Cloud customers get up to $300 in free credits |
| Web | ||
| Made by | Microsoft | Google LLC |
| Price read from | azure.microsoft.com | cloud.google.com |
| Checked | Nov 12, 2026 | Nov 12, 2026 |
Worth knowing
Azure AI Speech
- Prices follow the region picked on the pricing page; Neural is about $15 per 1M characters and Neural HD about $22 in the US regions.
- Custom voice is limited-access, and training and endpoint hosting are charged separately on top of synthesis.
- Prices are described as estimates and can vary with the agreement type and currency.
- The product was renamed from Azure AI Speech to Azure Speech in Foundry Tools; older docs and tutorials use the earlier name.
Google Cloud Text-to-Speech
- Billing is per character after the free allowance: US$4 per 1M for WaveNet and Standard, US$16 for Neural2, US$30 for Chirp 3 HD and US$160 for Studio.
- Instant custom voice has no free allowance and costs US$60 per 1M characters.
- Gemini-TTS models are billed per token, and the Gemini 3.8 models are Preview with rates that rise on January 1, 2027.
- It is an API and console service with no editor for non-developers, and billing must be enabled before any use.
Key features
Azure AI Speech
- Neural and Neural HD voices
- Prebuilt text-to-speech voices come in Neural and Neural HD variants, plus a MAI-Voice model priced alongside them.
- Custom and personal voice
- Professional custom voices are trained and hosted on Azure. Personal voice has free voice creation. Custom voice is a limited-access feature.
- Speech to text and translation
- The same service transcribes calls and meetings and translates speech, with captioning in more than 100 languages.
- Avatars
- Text-to-speech avatars (standard, photo and video) are billed per minute as an add-on to speech synthesis.
- Voice Live API
- A real-time voice agent API billed per million tokens, with pricing tiers set by the underlying language model.
Google Cloud Text-to-Speech
- Several voice generations
- Standard, WaveNet, Neural2, Studio, Polyglot and Chirp 3 HD voices sit at different price points. Google cites 380+ voices across 75+ languages and variants.
- Chirp 3 HD voices
- Conversational voices built for agents, with low-latency streaming and spontaneous-sounding delivery.
- Gemini-TTS
- Style, accent, pace, tone and emotion are steered with plain-language prompts, including multi-speaker speech.
- Instant custom voice
- Chirp 3 instant custom voice builds a personal voice model from as little as 10 seconds of audio, in 30+ locales.
- SSML and plain-text control
- Pauses, number and date formatting, pronunciation and delivery are set with SSML tags or plain scripting, depending on the voice model.
Plans
Azure AI Speech
Free (F0)$0
0.5M Neural characters a month for text to speech, 5 audio hours for speech to text and translation.
Google Cloud Text-to-Speech
Free monthly allowance$0
4M characters on Standard and WaveNet, 1M on Neural2, Studio, Polyglot and Chirp 3 HD. Billing must be enabled.
Common questions
Which is cheaper, Azure AI Speech or Google Cloud Text-to-Speech?
As of Nov 12, 2026, Azure AI Speech publishes no starting price, and Google Cloud Text-to-Speech publishes no starting price.
Which has a free plan, Azure AI Speech or Google Cloud Text-to-Speech?
As of Nov 12, 2026, both Azure AI Speech and Google Cloud Text-to-Speech have a free plan.
More comparisons
How this page was made
- The verdict at the top follows a fixed rule. When both tools are in the same category and either is in our top three for it, the verdict follows that ranking: the higher-placed tool is our pick. The top three weigh who each tool suits as well as its grades, so a pick can go against the GPA. Otherwise the page does not pick one and says which grades higher: the higher GPA first; on a tie, the one with a free plan, then the one that starts cheaper in dollars a month, counted the same way for both; if they are still level, what the choice depends on.
- The grades are our own, from each tool’s review or report card. The GPA is their average.
- Prices, plans and platforms come from each tool’s listing. The table above says where each price was read and when.
- The differences and the answers to the questions are worked out from those listings.