Benchmark Almanac

DesignArena Text-to-Speech

Google Gemini 3.8 Flash TTS 1,468 tops its leaderboard Grok TTS 1,226 leads the models we track
Not in model cards

Speech models read the same text aloud. Users listen to two anonymous clips and vote for the more natural one.

One of 40 DesignArena boards. DesignArena ranks entries with Bradley-Terry ratings from user votes.

Kind
Audio understanding and voice
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Last model added
October 2026
Scores refreshed
November 12th, 2026
36
Entries on its own leaderboard, 4 models we track

Google Gemini 3.8 Flash TTS tops the leaderboard with 1,468

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Grok TTS, at 1,226.

36 entries. Scroll sideways for the rest. Bars start at 800.

  1. 1,468 Google Gemini 3.8 Flash TTS
  2. 1,428 Google Gemini 3.1 Flash TTS
  3. 1,384 Google Gemini 2.5 Pro TTS
  4. 1,363 Eleven v4 Turbo
  5. 1,353 Eleven v4
  6. 1,341 Google Gemini 2.5 Flash TTS
  7. 1,262 Eleven v3
  8. 1,226 Grok TTS
  9. 1,224 Cartesia Sonic 3.6
  10. 1,202 OpenAI GPT Realtime 2
  11. 1,188 GPT-4o Mini TTS
  12. 1,180 OpenAI GPT Realtime 2.1
  13. 1,166 MAI-Voice-2
  14. 1,150 Hume Octave 2
  15. 1,150 Voxtral Mini TTS 3
  16. 1,130 Inworld TTS-1.5 Max
  17. 1,125 MAI-Voice-2-Flash
  18. 1,123 Cartesia Sonic 3.5
  19. 1,122 Lightning v3.1 Pro
  20. 1,117 MiniMax Speech-2.5 Turbo
  21. 1,114 MiniMax Speech 2.8 HD
  22. 1,099 MiniMax Speech-02 HD
  23. 1,084 Speech 2.6
  24. 1,075 Murf AI Gen2
  25. 1,070 Lightning v3.1
  26. 1,017 Kalpa TTS Beta V0.1
  27. 1,013 Qwen3 TTS Flash
  28. 1,012 Kokoro TTS
  29. 995 Voxtral Mini TTS (2603)
  30. 992 Murf Falcon 2
  31. 986 Deepgram Aura v2
  32. 961 Amazon Polly
  33. 957 Silk Mulberry 1.5
  34. 947 Maya1
  35. 932 Chatterbox
  36. 869 Magpie TTS Multilingual

Bars start at 800.

  1. 1 Google Gemini 3.8 Flash TTS 1,468
  2. 2 Google Gemini 3.1 Flash TTS 1,428
  3. 3 Google Gemini 2.5 Pro TTS 1,384
  4. 4 Eleven v4 Turbo 1,363
  5. 5 Eleven v4 1,353
  6. 6 Google Gemini 2.5 Flash TTS 1,341
  7. 7 Eleven v3 1,262
  8. 8 Grok TTS 1,226
  9. 9 Cartesia Sonic 3.6 1,224
  10. 10 OpenAI GPT Realtime 2 1,202
  11. 11 GPT-4o Mini TTS 1,188
  12. 12 OpenAI GPT Realtime 2.1 1,180
  13. 13 MAI-Voice-2 1,166
  14. 14 Hume Octave 2 1,150
  15. 15 Voxtral Mini TTS 3 1,150
  16. 16 Inworld TTS-1.5 Max 1,130
  17. 17 MAI-Voice-2-Flash 1,125
  18. 18 Cartesia Sonic 3.5 1,123
  19. 19 Lightning v3.1 Pro 1,122
  20. 20 MiniMax Speech-2.5 Turbo 1,117
  21. 21 MiniMax Speech 2.8 HD 1,114
  22. 22 MiniMax Speech-02 HD 1,099
  23. 23 Speech 2.6 1,084
  24. 24 Murf AI Gen2 1,075
  25. 25 Lightning v3.1 1,070
  26. 26 Kalpa TTS Beta V0.1 1,017
  27. 27 Qwen3 TTS Flash 1,013
  28. 28 Kokoro TTS 1,012
  29. 29 Voxtral Mini TTS (2603) 995
  30. 30 Murf Falcon 2 992
  31. 31 Deepgram Aura v2 986
  32. 32 Amazon Polly 961
  33. 33 Silk Mulberry 1.5 957
  34. 34 Maya1 947
  35. 35 Chatterbox 932
  36. 36 Magpie TTS Multilingual 869

Source: Official DesignArena Text-to-Speech leaderboard

#EntryScoreSettingDateSource
1 Google Gemini 3.8 Flash TTS 1,468 Official leaderboard
2 Google Gemini 3.1 Flash TTS 1,428 Official leaderboard
3 Google Gemini 2.5 Pro TTS 1,384 Official leaderboard
4 Eleven v4 Turbo 1,363 Official leaderboard
5 Eleven v4 1,353 Official leaderboard
6 Google Gemini 2.5 Flash TTS 1,341 Official leaderboard
7 Eleven v3 1,262 Official leaderboard
8 Grok TTS 1,226 Apr 2026 Official leaderboard
9 Cartesia Sonic 3.6 1,224 Official leaderboard
10 OpenAI GPT Realtime 2 1,202 Official leaderboard
11 GPT-4o Mini TTS 1,188 Mar 2025 Official leaderboard
12 OpenAI GPT Realtime 2.1 1,180 Official leaderboard
13 MAI-Voice-2 1,166 Official leaderboard
14 Hume Octave 2 1,150 Official leaderboard
15 Voxtral Mini TTS 3 1,150 Oct 2026 Official leaderboard
16 Inworld TTS-1.5 Max 1,130 Official leaderboard
17 MAI-Voice-2-Flash 1,125 Official leaderboard
18 Cartesia Sonic 3.5 1,123 Official leaderboard
19 Lightning v3.1 Pro 1,122 Official leaderboard
20 MiniMax Speech-2.5 Turbo 1,117 Official leaderboard
21 MiniMax Speech 2.8 HD 1,114 Jan 2026 Official leaderboard
22 MiniMax Speech-02 HD 1,099 Official leaderboard
23 Speech 2.6 1,084 Official leaderboard
24 Murf AI Gen2 1,075 Official leaderboard
25 Lightning v3.1 1,070 Official leaderboard
26 Kalpa TTS Beta V0.1 1,017 Official leaderboard
27 Qwen3 TTS Flash 1,013 Official leaderboard
28 Kokoro TTS 1,012 Official leaderboard
29 Voxtral Mini TTS (2603) 995 Official leaderboard
30 Murf Falcon 2 992 Official leaderboard
31 Deepgram Aura v2 986 Official leaderboard
32 Amazon Polly 961 Official leaderboard
33 Silk Mulberry 1.5 957 Official leaderboard
34 Maya1 947 Official leaderboard
35 Chatterbox 932 Official leaderboard
36 Magpie TTS Multilingual 869 Official leaderboard

2 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI xAI
Own leaderboard 2 leaders
Grok TTS 1,226

Source: Official DesignArena Text-to-Speech leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard
1,100 1,150 1,200 1,250 2025 2026 2027 1,100 1,150 1,200 1,250 2025 2026 2027

Source: Official DesignArena Text-to-Speech leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Audio and video

All 233 in Audio and video
  • Video-MMMU In use 300 expert-level lecture videos with questions testing perception, comprehension and adaptation of knowledge. Accuracy. Qwen 3.8 Max 88.7%
  • LVBench In use Hours-long video understanding with multiple-choice questions on reasoning, grounding and summarization. Gemini 4 Argon 91.7%
  • MMVU In use 3,000 expert-annotated questions on videos across 27 disciplines. Multiple-choice and open-ended accuracy. Qwen 3.8 Max 82.4%
  • MLVU In use Multi-task long video understanding (minutes to hours) with multiple-choice and generation tasks. Accuracy. Qwen 3.8 Max 90.8%
  • FLEURS In use Models transcribe read speech in 102 languages built from FLoRes-101 sentences. Model cards usually report word error rate, so lower scores are better. Phi-4-multimodal 4.00%
  • VoiceBench In use Spoken instructions with speaker, environment and content variation. Voice assistant responses scored by GPT judge or rules. Qwen3.5-Omni-Plus 93.1%