TTS per 1,000 characters: Sume $0.0475 vs MAI-Voice-2.1 $0.022

Sume text to speech costs $0.0475 per 1,000 characters. Microsoft's page lists MAI-Voice-2.1 at $22 per 1M characters, which is $0.022 per 1,000, and MAI-Voice-2.1-Flash at $15 per 1M, which is $0.015 per 1,000 (read 2026-10-08). On list price alone, Microsoft's numbers are lower; the Sume number is the provider list price times 1.25.

This is a list-price comparison and nothing more. The two are different products with different voices, delivery and limits, and I make no quality claim either way. Use it to size a budget, not to pick a voice.

Price at four volumes

One Sume request takes at most 20,000 characters, so the larger rows are several jobs. The Sume column is characters times $0.0475 per 1,000.

Where the numbers come from

The Sume rate is the catalog rate: one provider credit per character at the provider list, times the 1.25 Sume multiplier, reserved at submit. A shortfall in the workspace balance returns 402 insufficient_credits before any audio is made. Spaces and punctuation count as characters.

The Microsoft figures are from Microsoft AI's announcement, which also states 150 ms for 45 s of audio and 23 languages across 26 locales for MAI-Voice-2.1. Sume TTS is an asynchronous job you poll or receive by webhook, so a latency figure for a streaming product does not map onto it.

When the gap matters

Per-character cost is rarely the whole bill. A 60-second voiceover is roughly 900 characters, so the difference between the two lists is a couple of cents per video. It starts to matter at scale:

• A 900-character voiceover: $0.04275 on Sume, $0.0198 at the MAI-Voice-2.1 list.
• A thousand of those a month: $42.75 versus $19.80.
• Compare the full pipeline, including assembly and captions, before treating the voice line as the cost driver.