Flash v2.5 text-to-speech model
ElevenLabsReleased Dec 2024.
Flash v2.5 is a text-to-speech (TTS) model from ElevenLabs. It ranks #55 of 103 on the Artificial Analysis TTS arena with Elo 1075 (95% CI ±10, 5,849 samples). On TTS Arena V2 it scores Elo 1494 (±23, 1,214 votes). List price is about $50 per 1M characters. Released Dec 2024.
Key facts
- Provider
- ElevenLabs
- AA rank
- #55 / 103
- AA Elo (95% CI)
- 1075 ±10
- AA samples
- 5,849
- TTS Arena Elo (uncertainty)
- 1494 ±23
- TTS Arena votes
- 1,214
- TTS Arena win rate
- 51%
- Price (per 1M chars)
- $50
- Released
- Dec 2024
- Voices tested by AA
- 7
Not yet available: Open weights, License, Languages, Voice cloning
Who it is for (from the data)
- When quality is not the top priority: #55 of 103, in the bottom half; if quality matters most, check higher-ranked models first.
- Larger budgets: $50 per 1M characters, in the most expensive third of 83 priced models.
- Use via ElevenLabs's service or API; the sources do not list it as open weights.
- Existing ElevenLabs users: this vendor has 7 models on the board.
Nearby models
By AA Elo; this model highlighted| # | Model | AA Elo | Arena Elo | Price $/1M chars | Open weights | Δ |
|---|---|---|---|---|---|---|
| 52 | 1086 | – | $15 | – | +11 | |
| 53 | 1085 | 1487 | $15 | – | +10 | |
| 54 | 1083 | – | $9.2 | – | +8 | |
| 55 | 1075 | 1494 | $50 | – | ||
| 56 | 1072 | – | $50 | – | -3 | |
| 57 | 1072 | – | $191.6 | – | -3 | |
| 58 | 1072 | – | $160 | – | -3 |
Alternatives · Cheaper with AA Elo ≥ 1045
| # | Model | AA Elo | Arena Elo | Price $/1M chars | Open weights |
|---|---|---|---|---|---|
| 2 | 1331 | – | $40 | – | |
| 3 | 1301 | – | $19.3 | – | |
| 4 | CSonic 3.6 Cartesia | 1293 | – | $49 | – |
| 5 | 1282 | – | $16.5 | – | |
| 8 | 1266 | – | $19.3 | – |
Alternatives · More from ElevenLabs
| # | Model | AA Elo | Arena Elo | Price $/1M chars | Open weights |
|---|---|---|---|---|---|
| 1 | 1334 | – | $80 | – | |
| 2 | 1331 | – | $40 | – | |
| 18 | 1212 | – | $50 | – | |
| 23 | 1187 | 1499 | $100 | – | |
| 48 | 1092 | 1493 | $50 | – |
Alternatives · Top 5 open-weight by AA Elo
| # | Model | AA Elo | Arena Elo | Price $/1M chars | Open weights |
|---|---|---|---|---|---|
| 12 | 1239 | – | $15 | Open | |
| 13 | BBreeze TTS 2 BreezeBlue | 1226 | – | $34 | Open |
| 22 | 1189 | – | – | Open | |
| 42 | 1108 | – | $15 | Open | |
| 46 | 1097 | – | – | Open |
How to read the scores
- AA Elo: the Elo score from the Artificial Analysis speech arena. Listeners blind-compare two clips and pick the better one; "±" is the 95% confidence interval.
- TTS Arena Elo: Elo from TTS Arena V2 community blind votes. The two boards use different scales, so do not subtract one from the other.
- If the gap between two models is smaller than their combined intervals, the order is not statistically reliable; treat them as the same tier. More samples or votes mean narrower intervals.
FAQ
Where does Flash v2.5 rank among TTS models?
As of Nov 7, 2026, Flash v2.5 ranks #55 of 103 on the Artificial Analysis TTS leaderboard with Elo 1075 (±10).
Is Flash v2.5 open source?
The sources do not list Flash v2.5 as open-weight; it is generally used through the ElevenLabs API.
Which top-10 TTS model is cheapest?
Among the AA top 10, the lowest list price is Gemini 3.8 Flash-Lite TTS (Google) at $11 per 1M characters, Elo 1254.
What is the best open-source TTS model?
By AA Elo, the top open-weight model is Fish Audio Drama 3 Preview by Fish Audio (Elo 1239, #12 overall); 18 open-weight models have an AA score.
How much does Flash v2.5 cost?
According to Artificial Analysis, Flash v2.5 costs $50 per 1M characters. Prices change; check the vendor for current pricing.
Who makes Flash v2.5?
Flash v2.5 is made by ElevenLabs; Artificial Analysis lists its release as Dec 2024.
How reliable is Flash v2.5's score?
The scores are based on 5,849 Artificial Analysis samples (95% CI ±10), 1,214 TTS Arena votes. More samples and narrower intervals mean a more stable rank.
How far is Flash v2.5 from #1?
#1 on Artificial Analysis is Eleven v4 (Elo 1334); Flash v2.5 has Elo 1075, 259 points behind.
Sources
- Artificial Analysis TTS leaderboard ↗
- TTS Arena V2 (TTS-AGI, Hugging Face) ↗
- Model page (as listed on TTS Arena) ↗
17NAS syncs the upstream leaderboards daily; this page's data was synced on Nov 7, 2026.
Price and release date come from Artificial Analysis, recorded on Oct 25, 2026.
"–" means the upstream sources do not provide this value; we do not guess.