Qwen3-TTS is cheaper: $0.0060 vs $0.083 per 1,000 characters on each model's cheapest path — 13.9× less. For 10,000 characters (~12 min of audio): $0.060 vs $0.833. Stamped 2026-08-28. ElevenLabs Flash v2.5's premium may buy quality — this index measures cost only.
Qwen3-TTS vs ElevenLabs Flash v2.5: price
Qwen3-TTS
$0.0060–$0.09 per 1,000 characters
Cheapest: $0.0060 via Self-hosted (rented GPU) (estimated)
ElevenLabs Flash v2.5
$0.083 per 1,000 characters
Cheapest: $0.083 via ElevenLabs (Starter tier (cheapest incl.)) (derived)
How much does Qwen3-TTS cost?
| Provider | Price /1k chars | Provider's own unit | Status | Source |
|---|---|---|---|---|
| fal.ai · 1.7B | $0.09$0.076/min of audio | $0.00009 per characterUSD per character | Derived2026-08-28 | source |
| fal.ai · 0.6B | $0.07$0.059/min of audio | $0.00007 per characterUSD per character | Derived2026-08-28 | source |
| Self-hosted (rented GPU) (self-host)cheapest | $0.0060$0.0051/min of audio | rented GPU timeGPU-hours | Estimated2026-08-28 | source |
| Alibaba Cloud (DashScope) | DashScope console ratesUSD per character | Hidden | source |
How these numbers were computed
- fal.ai (1.7B): $0.00009/char (fal.ai listed rate, captured in FairStack operational cost audit, 2026-03-23 (price-audit-full.json)) × 1,000 = $0.09/1k chars.
- fal.ai (0.6B): $0.00007/char (FairStack operational cost audit, 2026-03-23 (price-audit-full.json)) × 1,000 = $0.07/1k chars.
- Self-hosted (rented GPU): Open weights on commodity GPU rental; operator-measured throughput ≈ $0.006/1k chars excluding setup and idle time.
How much does ElevenLabs Flash v2.5 cost?
| Provider | Price /1k chars | Provider's own unit | Status | Source |
|---|---|---|---|---|
| ElevenLabs · Starter tier (cheapest incl.)cheapest | $0.083$0.071/min of audio | 0.5 credits per charactercredits | Derived2026-08-28 | source |
How these numbers were computed
- ElevenLabs (Starter tier (cheapest incl.)): Starter $5/mo ÷ 30,000 credits × 0.5 credit/char × 1,000 = $0.083/1k chars.
Frequently asked
Is Qwen3-TTS cheaper than ElevenLabs Flash v2.5?
Qwen3-TTS is cheaper: $0.0060 vs $0.083 per 1,000 characters on each model's cheapest path — 13.9× less. For 10,000 characters (~12 min of audio): $0.060 vs $0.833. Stamped 2026-08-28. ElevenLabs Flash v2.5's premium may buy quality — this index measures cost only.
What does 10,000 characters (~12 min of audio) cost on each?
Qwen3-TTS: $0.060 (cheapest path). ElevenLabs Flash v2.5: $0.833. Both computed from the normalized canonical rate; real invoices vary with resolution, minimum charges, and tiers.
Which is better quality?
This index measures cost, not quality. For quality benchmarks, see Artificial Analysis — a voice model that costs more is not automatically better, and vice versa.
Prices normalized to per 1,000 characters; per-clip, credit, and tier billing converted per themethodology. Quality comparison is out of scope — seeArtificial Analysis.