OpenAI TTS-1 HD 1106: high-definition text-to-speech via OrcaRouter's OpenAI-compatible API
The openai/tts-1-hd-1106 model is a text-to-speech model from OpenAI, specifically the high-definition variant released in November 2023. It converts text into natural-sounding audio with a focus on…
The openai/tts-1-hd-1106 model is designed for high-definition audio output. It generates natural-sounding speech with good prosody and emotion. The model supports multiple voices (as provided by OpenAI) and allows adjustment of speed and sample rate. Output is typically 24kHz or 48kHz audio depending on configuration. This makes it suitable for professional applications like media production.
Optimal uses include generating voiceovers for videos, creating narration for audiobooks, building voice interfaces in applications, and adding speech output to assistive technologies. The high-definition quality is especially valuable when output will be heard by end users in customer-facing scenarios. For internal testing or low-fidelity prototypes, consider lower-cost alternatives.
If your use case does not require high-fidelity audio—for example, simple alert tones, very short utterances, or internal logging—a lower-cost TTS model may be more economical. The $30 per 1M token pricing applies to both input and output, so generating many short clips can accumulate quickly. For production at scale, evaluate your quality requirements and budget.
Yes, OpenAI provides several default voices for this model (e.g., alloy, echo, fable, onyx, nova, shimmer). You can select a voice via API parameters. Additionally, you can adjust speed (0.5x to 2.0x), sample rate, and output format. The model does not support custom voice cloning or fine-tuning. OrcaRouter passes these parameters through to OpenAI without modification.
While specific benchmark scores for openai/tts-1-hd-1106 are not provided in the available data, it is widely recognized as OpenAI's highest-quality TTS model as of its release date (November 2023). It ranks among top models for naturalness and clarity in internal evaluations. For objective performance, consult OpenAI's documentation and third-party reviews.
Latency depends on the length of the input text and the selected audio format. In general, the model returns audio within a few seconds for short inputs (e.g., 100 characters). Longer texts may take proportionally longer. OrcaRouter does not add significant overhead beyond the provider's response time. Streaming can reduce perceived latency for real-time applications.
The model is limited to text-to-speech, without support for voice conversation or emotion control beyond what voice selection provides. It cannot generate background sounds or music. Output quality can degrade with unusual pronunciations, homonyms, or foreign words. Also, the model has a maximum input length (typically 4096 characters). For very long texts, segment input.
Pricing is $30.00 per 1 million input tokens and $30.00 per 1 million output tokens. Input tokens count the text you provide. Output tokens count the generated audio tokens. Both are billed at the same rate. There is no per-minute or per-character charge; tokenization is processed by OpenAI. OrcaRouter adds zero markup, so you pay exactly the published OpenAI rate.
The provided pricing is the standard rate through OrcaRouter. No volume discounts or cached pricing are mentioned in the available facts. You can reduce costs by optimizing input text length—shorter prompts generate fewer output tokens. For large-scale usage, consider negotiating directly with OpenAI, though OrcaRouter does not offer separate pricing tiers.
At $30 per 1M tokens for both input and output, this model is priced higher than many standard TTS models. For example, OpenAI's standard tts-1 model costs $15 per 1M input and $15 per 1M output. The HD variant commands a premium for higher quality. OrcaRouter passes through these provider rates without markup, so the cost difference is the same as using OpenAI directly.
Zero markup means OrcaRouter does not add any fees on top of the provider's per-token price. Your cost is exactly the OpenAI rate: $30 per 1M input and $30 per 1M output tokens. There are no platform surcharges, hidden costs, or usage thresholds. The only charges are those incurred by token usage at the published provider prices.
Use the OpenAI-compatible API at https://api.orcarouter.ai/v1. Set the model parameter to "openai/tts-1-hd-1106". Provide input text in the standard message format (e.g., role: user, content: your text). Additional TTS-specific parameters (like voice, speed, response_format) can be included in the request body. OrcaRouter forwards the request to OpenAI and returns the audio response. See OpenAI's TTS API documentation for full parameter details.
You can set the same parameters as the OpenAI TTS API: input (string), model ("openai/tts-1-hd-1106"), voice (string from available voices), speed (number 0.5–2.0), response_format (e.g., "mp3", "opus", "aac", "flac", "wav"), and sample_rate. Include them in the JSON body of a POST request to the chat/completions endpoint. OrcaRouter preserves all parameters and does not alter them.
Yes, you can use streaming by setting the stream parameter to true in your API request. The model will return audio chunks as they are generated, which is useful for real-time applications. OrcaRouter supports streaming without additional configuration. Ensure your client handles chunked responses appropriately, as standard for OpenAI-compatible streaming endpoints.
Replace your base API URL to https://api.orcarouter.ai/v1 and update the model identifier to "openai/tts-1-hd-1106". Your existing API key for OpenAI will not work; you need an OrcaRouter API key. The request body format remains identical. No other code changes are required. OrcaRouter's endpoint is designed to be a drop-in replacement for those familiar with the OpenAI SDK.
Both models are from OpenAI. The HD variant (tts-1-hd-1106) produces higher audio quality with more natural intonation and clarity compared to the standard tts-1 (priced at $15 per 1M tokens each way). The HD model costs double per token. The choice depends on your quality requirements; for casual use, the standard model may suffice. For professional production, the HD is recommended.
Compare to providers like Amazon Polly, Google Cloud Text-to-Speech, or Microsoft Azure Speech. The openai/tts-1-hd-1106 model is competitive in naturalness but typically priced higher per character. It excels in expressiveness and ease of integration via an OpenAI-compatible API. However, other providers offer more voices, language support, and custom voice creation. Evaluate based on your specific needs for language, quality, and budget.
Open-source models like Tacotron, FastSpeech, or Tortoise require setup and may not match the output quality of OpenAI's HD model. The hosted model offers convenience, reliability, and high quality without infrastructure management. However, open-source models can be cost-free for self-hosted use, though they incur compute costs. For small projects, open-source may be cheaper; for production requiring consistent quality, the HD model is a strong choice.
OpenAI-compatible — keep the SDK you already use
https://api.orcarouter.ai/v1response_formatspeedvoice| Input / 1M tokens | $30.00 |
| Output / 1M tokens | $30.00 |
| Currency | USD |
Estimate based on list price
Estimate only — actual token counts depend on the provider's tokenizer.
GET /api/public/models/openai/tts-1-hd-1106Open @misc{orcarouter_tts_1_hd_1106,
title = {openai/tts-1-hd-1106 API},
author = {openai},
year = {n.d.},
howpublished = {OrcaRouter},
url = {https://www.orcarouter.ai/models/openai/tts-1-hd-1106}
}openai. (n.d.). openai/tts-1-hd-1106 API. OrcaRouter. https://www.orcarouter.ai/models/openai/tts-1-hd-1106