OpenAI TTS-1 HD
Audio Deprecated Full comparison ↗ openai/tts-1-hd · by OpenAI
· neural-tts
Pricing — 1 offering(s)
About this offering: OpenAI TTS-1 HD (text-to-speech, high definition). Higher audio fidelity than TTS-1 at 2× the price. Same 6 preset voices and /v1/audio/speech endpoint. Priced at $30/1M characters. Deprecated: OpenAI removes TTS-1 HD (and TTS-1) from the API on January 6, 2027 and recommends gpt-realtime-2.1-mini instead, so plan new work on a different model.
Text characters
- $30.00 / 1M characters Current 2023-11-06 → 2027-01-05
Showing the active price and any recorded history. Full pricing history is available via the paid API — see API docs.
See how OpenAI TTS-1 HD fits into a cost-aware routing setup
See how →Capability profile
How OpenAI TTS-1 HD rates across core capability dimensions, with the task-level evidence behind each rating.
Noticeably higher audio fidelity than TTS-1, with better prosody and a less metallic sound. In the Artificial Analysis Text to Speech Arena (checked 2026-08-19), TTS-1 HD ranks #30 of 98 voices (Elo 1108), just ahead of TTS-1 (#34, Elo 1094).
Same 57-language support as TTS-1.
Same 6 preset voices as TTS-1; no cloning or custom voices.
Slightly higher latency than TTS-1 due to higher fidelity processing.
No voice cloning, no fine-tuning, no SSML. Same limitations as TTS-1.
Benchmarks
| Benchmark | Score | Config | Source |
|---|---|---|---|
| MOS (naturalness) | — | No independently published MOS. Vendor-described as higher fidelity than TTS-1; community rates it between TTS-1 and ElevenLabs Multilingual v2. | — |
Operator guidance
Deprecated: OpenAI removes TTS-1 HD from the API on January 6, 2027 (use gpt-realtime-2.1-mini), so don't build new long-term work on it. Until then, choose TTS-1 HD when output quality is more important than cost — the 2× price premium is justified for final production audio. For real-time apps and volume workloads, TTS-1 is sufficient. ElevenLabs Multilingual v2 exceeds TTS-1 HD in naturalness but costs 4× more.
Use cases
- Final narration and voice-over content where quality is paramount
- Podcast production and audio content where TTS-1 quality isn't sufficient
- Accessible content with premium audio requirements
Limitations
- Only 6 preset voices — same constraint as TTS-1
- No voice cloning or SSML support
- 2× cost of TTS-1; may be overkill for many use cases