Text-to-speech model optimized for speed.
TTS-1 is OpenAI's original production text-to-speech model, launched at OpenAI DevDay on November 6, 2023. It converts text into natural-sounding speech optimized for low-latency, realtime applications. The model is billed at $15.00 per million characters — half the price of its higher-fidelity sibling TTS-1 HD ($30.00/M chars) — making it the cost-efficient choice for streaming applications where perceived quality can be slightly reduced in exchange for faster first-audio delivery. It supports 9 built-in voices (alloy, ash, coral, echo, fable, onyx, nova, shimmer) and outputs audio in MP3, Opus, AAC, FLAC, WAV, or PCM formats.
As of 2025–2026, OpenAI positions gpt-4o-mini-tts as the recommended option for new voice-agent builds due to its instruction-following capabilities (e.g., 'speak like a sympathetic customer service agent'), while TTS-1 remains fully supported for teams that built on it and prioritise latency or cost predictability. TTS-1 is callable via TheRouter's OpenAI-compatible endpoint (/v1/audio/speech) with no client changes required — the same Authorization header and payload structure work unchanged. Rate limits are set per-tier: Free/Tier 1 at 3–500 RPM; Tier 5 at 10 000 RPM.
| Type | Rate |
|---|---|
| Text | $16.20 / 1M chars |
request is price per 1M characters
| Release date | 2023-11-06 (OpenAI DevDay)openai.com ↗ | verified |
| Pricing | $15.00 per 1 M charactersopenai.com/api/pricing ↗ | verified |
| Max input length | 4 096 characters per requestplatform.openai.com ↗ | verified |
| Supported voices | alloy, ash, coral, echo, fable, onyx, nova, shimmer (9 voices)platform.openai.com ↗ | verified |
| Output formats | MP3 (default), Opus, AAC, FLAC, WAV, PCMplatform.openai.com ↗ | verified |
| Speed control | 0.25–4.0× (default 1.0×)platform.openai.com ↗ | verified |
| Rate limits (Tier 1) | 500 RPMplatform.openai.com ↗ | verified |
| Instruction-following (voice style) | Not supported — use gpt-4o-mini-tts for prompt-controlled speechplatform.openai.com ↗ | verified |
| Training cutoff | Not publicly disclosed | unknown |
| License | Proprietary (OpenAI API Terms of Service)openai.com/policies ↗ | verified |
| Benchmark | Distribution | Score | Source |
|---|---|---|---|
Relative latency vs TTS-1 HD OpenAI states TTS-1 provides lower latency at the cost of lower quality versus TTS-1 HD. No numeric TTFA figures publicly disclosed. | Lower latency than TTS-1 HDqualitative | platform.openai.com ↗ |
Use the global api.therouter.ai endpoint shown below for new integrations; the legacy China accelerated endpoint is retired.
curl https://api.therouter.ai/v1/audio/speech -H "Content-Type: application/json" -H "Authorization: Bearer $THE_ROUTER_API_KEY" -d '{
"model": "openai/tts-1",
"input": "Welcome to TheRouter.",
"voice": "alloy"
}'curl https://api.therouter.ai/v1/audio/speech \
-H "Authorization: Bearer $THEROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/tts-1",
"input": "The quick brown fox jumped over the lazy dog.",
"voice": "alloy"
}' \
--output speech.mp3OpenAI released gpt-4o-mini-tts alongside new speech-to-text models, positioning it as the recommended model for voice agents. Unlike TTS-1, it accepts a free-text instructions field that controls accent, tone, emotion, and speaking style. TTS-1 remains available and is now the lower-latency, lower-cost legacy option.
Microsoft announced TTS-1 and TTS-1 HD voices on Azure OpenAI Service and Azure AI Speech, giving enterprise customers an Azure-native path to the same six original voices (alloy, echo, fable, onyx, nova, shimmer). This expanded TTS-1's reach to regulated industries where Azure data-residency is required.
TTS-1 is optimised for speed and low latency — ideal for real-time streaming and conversational applications. TTS-1 HD is optimised for audio quality and is better suited for pre-recorded content like podcasts, audiobooks, or any setting where a listener will scrutinise the output. TTS-1 costs $15/M characters; TTS-1 HD costs $30/M characters.
Migrate if you need instruction-following (controlling accent, tone, or personality via a free-text instructions field). Stick with TTS-1 if your pipeline is stable, you don't need style control, and you want predictable cost at $15/M characters. TTS-1 is not being deprecated and will continue to be supported.
Point your client's baseURL to https://api.therouter.ai/v1, use your TheRouter API key, and set model to openai/tts-1. All other parameters (voice, speed, response_format, input) remain identical to the OpenAI SDK. TheRouter handles routing, billing, and rate-limit aggregation transparently.
TTS-1 supports 9 built-in voices: alloy, ash, coral, echo, fable, onyx, nova, and shimmer. Custom or cloned voices are not supported by TTS-1. For zero-shot voice cloning, consider open-source models like fish-speech-1.5 or cosyvoice2-0.5b available on TheRouter.
| source | URL | retrieved | |
|---|---|---|---|
| Release date | openai.com ↗ | 2026-06-03 | verified |
| Pricing | openai.com/api/pricing ↗ | 2026-06-03 | verified |
| Max input length | platform.openai.com ↗ | 2026-06-03 | verified |
| Supported voices | platform.openai.com ↗ | 2026-06-03 | verified |
| Output formats | platform.openai.com ↗ | 2026-06-03 | verified |
| Speed control | platform.openai.com ↗ | 2026-06-03 | verified |
| Rate limits (Tier 1) | platform.openai.com ↗ | 2026-06-03 | verified |
| Instruction-following (voice style) | platform.openai.com ↗ | 2026-06-03 | verified |
| Training cutoff | — | — | unknown |
| License | openai.com/policies ↗ | 2026-06-03 | verified |
| Relative latency vs TTS-1 HD | platform.openai.com ↗ | 2026-06-03 | verified |
| OpenAI launches gpt-4o-mini-tts with instruction-following for voice agents | openai.com ↗ | 2026-06-03 | verified |
| OpenAI TTS voices arrive on Azure OpenAI Service | techcommunity.microsoft.com ↗ | 2026-06-03 | verified |
| What is the difference between TTS-1 and TTS-1 HD? | platform.openai.com ↗ | 2026-06-03 | to verify |
| Should I migrate from TTS-1 to gpt-4o-mini-tts? | platform.openai.com ↗ | 2026-06-03 | to verify |
| How do I use TTS-1 via TheRouter? | platform.openai.com ↗ | 2026-06-03 | to verify |
| What voices are available, and can I use custom voices? | platform.openai.com ↗ | 2026-06-03 | to verify |