CPU-only TTS results for cheaper voice AI agents
Three TTS models were compared using only a CPU, with no GPU. The test used an Intel Xeon CPU with 4 cores and 15.6GB of memory, and it tried six text lengths from 12 to 1712 characters. Each setup was timed 5 times after warmup, for 150 timed runs in total, and every audio sample was scored with UTMOS.
Inflect-Nano-v1 was the fastest at RTF 0.1376 and MOS 3.48, but the score may be too generous because the voice sounded buzzy, metallic, and flat. Supertonic-3 in 2-step mode reached RTF 0.1781 and MOS 1.53, making it fast but low quality. Supertonic-3 in 5-step mode reached RTF 0.3164 and MOS 4.37, giving a stronger balance of speed and quality.
Kokoro-82M reached MOS 4.44 with ONNX and MOS 4.45 with , but it was slower at RTF 0.5711 and 0.7865. Inflect-Nano-v1 also has an output cap of about 15 seconds, so it can silently cut off longer text and make long-text speed numbers look better than they really are.
Key points
- Three TTS models were tested on an Intel Xeon 4-core CPU with 15.6GB of memory.
- Inflect-Nano-v1 was fastest at RTF 0.1376, but it may cut off audio after about 15 seconds.
- Supertonic-3 5-step reached MOS 4.37 and was faster than Kokoro-82M.
- Kokoro-82M had the highest quality scores, with MOS 4.44 to 4.45, but it ran more slowly.
- UTMOS may overrate some small -based models, so listening tests still matter.