Deepgram sells more than one text to speech model, at different prices, and only one of them is the model Deepgram tells new builds to use. Flux TTS, the current one, costs $45 per million characters on pay as you go, which is more than Bland Speech's $40. Aura-2, a previous-generation model by Deepgram's own description, costs $30.
So a team comparing the two companies is really comparing Bland Speech with two different Deepgram products, one cheaper and one newer, and the answer changes depending on which Deepgram voice sits on the other side of the table. Both stream audio over a WebSocket, both return phone-ready audio, and the real split is how each voice sounds to a caller, how each is priced, and where a team wants the model to run.
Which is better, Bland Speech or Deepgram Aura?
For a voice agent on a phone line where callers should hear a person, Bland Speech is the better choice as of November 2026. Bland Speech v3 ranked first of 13 AI voices in Design Arena's Audio Realism Bench results of 4 August 2026, and it bills one flat rate of $0.04 per 1,000 characters with no plan to choose.
Deepgram makes more sense for a team that wants the lowest price per character, the lowest published first-audio figure, or a model it can run self-hosted. Deepgram wasn't in those 4 August results, so those results don't rank the two voices against each other.
Use Bland Speech when
- A voice agent takes phone calls where a polished, read-aloud voice sounds fake
- A team wants one per-character rate, with no plan or credit tier to choose
- An existing OpenAI audio speech integration should switch providers by changing a base URL, a key, and a model name
- A script needs performed delivery, such as a laugh or a warm read, written inline as tags
- A team wants an instant voice clone from about 10 seconds of audio
Use Deepgram when
- A team wants a lower price per character and will run Aura-2, a previous-generation model, at $30 per million
- A product needs many simultaneous streams on a self-serve plan, where Deepgram publishes up to 45 streaming connections in North America on pay as you go
- A security team wants to run text to speech in a private cloud or on-premise under Deepgram's terms
- A team wants speech to text and text to speech from one vendor, with guides for Twilio, Pipecat, and LiveKit
What these rows mean
- Audio Realism Bench, 4 August 2026: Audio Realism Bench is Design Arena's blind listening test, where native speakers hear two clips of the same line from two models and pick the one that sounds more like a real person. A higher Elo score means listeners picked that voice more often. Bland Speech v3 scored 1365 in those results, and Deepgram's voices were not among them.
- Price per million characters: Bland Speech costs $0.04 per 1,000 characters, so $40 per million, with no subscription. Deepgram bills per 1,000 characters of input text: $0.045 for Flux TTS, $0.030 for Aura-2, and $0.015 for Aura-1 on pay as you go.
- Free usage: New Bland accounts get 50,000 characters free, nearly an hour of speech. Deepgram's Free plan is $200 of credit with no card and no expiry.
- First audio: Each figure is the company's own published claim and neither is a race Bland ran. Bland publishes 199ms to first audio on a streamed request, and Deepgram says Flux TTS starts in "as low as 80ms at full 24kHz fidelity, even under production load".
- Phone audio: US phone networks and Twilio carry mu-law at 8 kHz, and both companies return it directly.
- Delivery controls: Bland Speech reads inline tags such as [laughs], [sighs], and [say warmly]. Deepgram's Flux TTS has a delivery dial from -2 to 2 and a speed range of 0.5 to 1.5.
How do Bland Speech and Deepgram compare?
The two companies built different things. Bland Speech is a text to speech model made for phone calls, where, in Bland's words, polish reads as fake. Deepgram built Flux TTS to stream turn after turn for voice agents, and says it keeps tone across turns.
Bland Speech overview
Bland Speech is the text to speech model inside Bland, a voice AI provider that builds and runs its own language, speech, and transcription models for enterprise phone calls. Bland Speech is sold on its own as an API and a studio on the Bland Speech page.
A team sends text over HTTP to POST /v2/tts and gets raw audio or a WAV file back, or streams language model tokens into a WebSocket at WSS /v2/tts/ws, where a new context interrupts a turn that is playing. POST /v2/audio/speech follows the OpenAI audio speech format, and a first-party Pipecat service and a LiveKit Agents plugin for Python are available. The text to speech API page has the code.
Pros
- Bland Speech v3 ranked first of 13 AI voices on Audio Realism Bench on 4 August 2026
- One flat rate, $0.04 per 1,000 characters, with no monthly plan
- An OpenAI-compatible endpoint, so an existing integration moves over by changing a base URL
- Performance tags and a director in the studio for how a line is read
- Mu-law at 8 kHz and PCM from 8 to 48 kHz, ready for phone calls
- Instant clones from about 10 seconds of audio, and professional clones from 30 minutes or more
Cons
- More than a quick prototype needs, because Bland Speech is built for production calls
Deepgram overview
Deepgram sells three generations of text to speech. Flux TTS runs on a /v2/speak endpoint over a WebSocket or REST, and Deepgram says it is "recommended for all new builds". Aura-2 and Aura sit on /v1/speak and are, in Deepgram's words, the previous and first generations. Aura-2 has voices in seven languages and Aura is English only.
Flux TTS speaks English plus 10 languages with built-in text normalization, and Deepgram says the same model handles 40 or more additional languages for teams that bring their own normalization. Deepgram says Flux TTS can run self-hosted, in a private cloud, or on-premise.
Pros
- Flux TTS first audio as low as 80ms, by Deepgram's own claim
- Aura-2 at $30 per million characters and Aura-1 at $15 per million on pay as you go
- $200 of free credit with no card and no expiry
- Mu-law and A-law output for phone lines
- Private cloud and on-premise deployment for Flux TTS
- Streaming connections up to 45 in North America on pay as you go, and up to 60 on Growth
Cons
- Flux TTS costs $45 per million characters, above Bland Speech's $40
- Aura-2, the cheaper model, is previous generation by Deepgram's own description
- Deepgram marks 32 of Flux TTS's 42 language codes as Early Access
- The delivery dial is in beta, and SSML is stripped rather than applied
- The best per-character rates on Aura models need a Growth plan of $4,000 or more a year
How do Bland Speech and Deepgram handle phone audio and latency?
Both return audio a phone line can carry without conversion, and on published numbers Deepgram is faster. Bland publishes 199ms to first audio on a streamed request, and Deepgram says Flux TTS responds in as low as 80ms. A team where every millisecond of a voice agent's turn counts has a real reason to pick Deepgram.
Bland Speech outputs mu-law at 8 kHz and PCM at 8, 16, 24, 44.1, or 48 kHz. Deepgram returns mu-law and A-law at 8 kHz or 16 kHz from its WebSocket, and a Deepgram guide covers Twilio, where mu-law at 8 kHz matches Twilio's Media Streams format. Bland Speech has no documented Media Streams path and is not one of Twilio ConversationRelay's voice providers, so a team building on Twilio's own tools has an easier route with Deepgram, while a team on Pipecat or LiveKit has a plugin for either.
How do Bland Speech and Deepgram handle security?
Deepgram says it holds SOC 2 Type 1 and Type 2, is HIPAA compliant, and signs business associate agreements (BAAs) for Enterprise customers handling electronic protected health information. Deepgram also offers a dedicated EU endpoint.
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. Bland's trust page says speech synthesis runs inside the FedRAMP boundary and that Bland runs its own model and speech instead of wrapping a third party. Enterprise customers get dedicated orchestration and dedicated GPUs run by Bland.
How much do Bland Speech and Deepgram cost?
Bland Speech and Deepgram land within a few dollars of each other per million characters, and which one costs less depends on the Deepgram model. Bland Speech costs $40 per million, Deepgram's Flux TTS costs $45, and Aura-2 costs $30.
Bland Speech pricing
Bland Speech costs $0.04 per 1,000 characters, directed or not, in the studio or the API. New accounts get 50,000 characters free, and credits never expire. A 60-second note costs about $0.04, a 10-minute narration about $0.36, and a million characters $40.
Enterprise pricing is custom, set with Bland's sales team.
Deepgram pricing
Deepgram bills per 1,000 characters of input text. On pay as you go, which has no minimum, Flux TTS is $0.045, Aura-2 is $0.030, and Aura-1 is $0.015. Growth is prepaid credit from $4,000 a year, where Aura-2 drops to $0.027 and Aura-1 to $0.0135 while Flux TTS stays at $0.045. Growth usage past the prepaid credit bills at the Growth rate plus 10%. Enterprise pricing is custom.
What does a month of speech cost on each?
At 100,000 characters a month, about two hours of speech, Bland Speech costs $4, Deepgram's Flux TTS costs $4.50, and Aura-2 costs $3. At 1 million characters, about 20 hours, the figures are $40, $45, and $30. At 10 million characters, about 200 hours, they are $400, $450, and $300 on pay as you go.
Deepgram's $200 free credit covers roughly 4.4 million characters on Flux TTS, against Bland's one-time 50,000, so a team testing at volume has more room to experiment on Deepgram.
Which sounds more human, Bland Speech or Deepgram Aura?
Design Arena's results don't rank the two directly, because Deepgram wasn't among the 13 AI voices scored on 4 August 2026. What those results show is that Bland Speech v3 ranked first of the 13 at 1365, against 1500 for a human recording.
- #1BlandBland Speech v31365
- 2ndMicrosoftMAI-Voice-21224
- 3rdxAIGrok TTS1156
- 4thOpenAIGPT Realtime 21119
- 5thGoogleGemini 2.5 Pro TTS1109
- Not rankedDeepgram's own voicesNot ranked
Ranks are among 13 AI voices, and Deepgram's own voices weren't ranked. Real human recordings scored 1500, at the right edge of this scale. Bars start at 600, not zero. Source: Design Arena's published results.
Design Arena's own words put Bland Speech v3 at an Elo of 1365, Microsoft's MAI-Voice-2 at 1224, and xAI's Grok TTS at 1156. A team choosing between Bland Speech and Deepgram can settle the question for its own use case by playing both voices the same script in a phone-quality format. Design Arena's published results explain how the votes were collected.
Why is Bland Speech the best Deepgram Aura alternative?
Bland Speech is the better Deepgram Aura alternative because callers hear a voice that ranked first among AI voices in an independent blind test on 4 August 2026, and because one flat rate replaces a choice between model generations and plans.
A team already on OpenAI's audio speech API can move by changing a base URL, and a team that wants a brand voice can clone one from about 10 seconds of audio, or train a professional clone on 30 minutes or more after Bland's consent and rights review. Every clone stays private to the organization that made it.
How was this comparison made?
Deepgram details come from Deepgram's own pricing, product, and developer pages, and voice scores come from Design Arena's published results of 4 August 2026.
Bland's comparison of text to speech APIs covers more vendors, and the head-to-head comparisons collect the pairs. A team that wants Deepgram on its own can read the Deepgram Aura review or the Deepgram Aura alternatives.
Frequently asked questions
Is Bland Speech better than Deepgram?
For phone calls where callers should hear a person, Bland Speech is better than Deepgram. Bland Speech v3 ranked first of 13 AI voices on Audio Realism Bench on 4 August 2026, and Deepgram's voices weren't in those results. Deepgram suits a team that wants the lowest published first-audio figure or a self-hosted model.
Is Bland Speech cheaper than Deepgram?
Against Deepgram's current Flux TTS, yes: $40 per million characters against $45. Against Aura-2, no: Deepgram charges $30 per million on pay as you go. Aura-1 is cheaper still at $15 per million, and Aura-2 is a previous-generation model by Deepgram's own description.
Which is faster, Bland Speech or Deepgram?
On each company's published claim, Deepgram. Bland publishes 199ms to first audio on a streamed request, and Deepgram says Flux TTS starts in as low as 80ms. Neither number comes from a race run by Bland.
What is the difference between Deepgram Aura and Flux TTS?
Flux TTS is Deepgram's current, conversation-first model on the /v2/speak endpoint, and Deepgram recommends it for new builds. Aura-2 and Aura are earlier generations on /v1/speak, cheaper per character, with seven languages on Aura-2 and English only on Aura.
Does Deepgram work with Twilio, Pipecat, and LiveKit?
Yes. Deepgram has guides for all three, and says requesting mu-law at 8 kHz matches Twilio's Media Streams format. Bland Speech has a first-party Pipecat service and a LiveKit plugin for Python, and is not a Twilio ConversationRelay voice provider.
Can Bland Speech replace a Deepgram text to speech integration?
A team on Pipecat or LiveKit can swap in Bland Speech through a plugin or service. A team on Deepgram's own API points its calls at POST /v2/tts or the WebSocket at /v2/tts/ws, and an integration written for OpenAI's audio speech format needs only a new base URL, a key, and a model name.
Does Bland Speech work on phone calls?
Yes. Bland built Bland Speech for phone calls, and it returns mu-law at 8 kHz as well as PCM from 8 to 48 kHz, so audio goes onto a phone line without conversion.
