Alibaba AI voice model cracks top 5 globally, outperforming US rivals in regional accentsThe new model supports more than 30 languages, seven major Chinese dialects and over 20 regional accents2-MIN READ2-MIN0ListenPublished: 7:00pm, 29 May 2026A new Artificial Intelligence voice model from Alibaba-group-holding" class="entity-link entity-organization" data-entity-id="16108" data-entity-type="organization">Alibaba Group Holding has beaten out Western rivals OpenAI and xAI on a major global benchmark, underscoring its technical edge in capturing complex Chinese dialects and accents.Fun-Realtime-TTS-Preview, developed by Alibaba’s Tongyi Lab, has secured the fifth spot on the Artificial Analysis Speech Arena leaderboard with a score of 1,190. It was the only Chinese-engineered voice system in the global top five.Alibaba owns the China-morning-post" class="entity-link entity-organization" data-entity-id="12558" data-entity-type="organization">South China Morning Post.Select VoiceSelect Speed0.8x0.9x1.0x1.1x1.2x1.5x1.75x00:0000:001.00x
SRCSouth China Morning Post
LANGEN
LEANCenter-Right
WORDS110
ENT10
FRI · 2026-05-29 · 11:00 GMTBRIEF NSR-2026-0529-80166
NSR-2026-0529-80166News Report·EN·Technology
Alibaba AI voice model cracks top 5 globally, outperforming US rivals in regional accents
Alibaba's new AI voice model, Fun-Realtime-TTS-Preview, developed by Tongyi Lab, has achieved fifth place globally on the Artificial Analysis Speech Arena leaderboard with a score of 1,190. This marks the first time a Chinese-engineered voice system has entered the top five, outperforming Western competitors like OpenAI and xAI.
Minxiao ChangSouth China Morning PostFiled 2026-05-29 · 11:00 GMTLean · Center-RightRead · 1 min
South China Morning PostFIG 01
Reading time
1min
Word count
110words
Sources cited
0cited
Entities identified
10entities
Quality score
100%
§ 01
Briefing Summary
AI-generatedNEWSAR · AI
Alibaba's new AI voice model, Fun-Realtime-TTS-Preview, developed by Tongyi Lab, has achieved fifth place globally on the Artificial Analysis Speech Arena leaderboard with a score of 1,190. This marks the first time a Chinese-engineered voice system has entered the top five, outperforming Western competitors like OpenAI and xAI. The model demonstrates a particular strength in accurately capturing complex Chinese dialects and regional accents, supporting over 30 languages, seven major Chinese dialects, and more than 20 regional accents. This achievement highlights Alibaba's technical advancements in voice AI, especially in understanding diverse linguistic variations.
Confidence 0.85Claims 4Entities 10
§ 02
Article analysis
Model · rule-basedFraming
Technology
Economic Impact
Tone
Measured
AI-assessed
CalmNeutralAlarmist
Factuality
0.90 / 1.00
Factual
LowHigh
Sources cited
0
No named sources
FewMany
§ 03
Key claims
4 extracted01
Fun-Realtime-TTS-Preview is the only Chinese-engineered voice system in the global top five.
factual
Confidence
1.00
02
The model supports more than 30 languages, seven major Chinese dialects, and over 20 regional accents.
factual
Confidence
1.00
03
The Alibaba AI voice model outperformed US rivals OpenAI and xAI on a major global benchmark.
factual
Confidence
1.00
04
Alibaba AI voice model Fun-Realtime-TTS-Preview secured fifth spot on Artificial Analysis Speech Arena leaderboard with a score of 1,190.
statisticAlibaba’s Tongyi Lab
Confidence
1.00
§ 04
Full report
1 min read · 110 words§ 05
Entities
10 identifiedKey playerOppositionContextPositiveNeutralNegative
Organizations6
§ 06
Keywords & salience
10 termsalibaba ai voice model
1.00
artificial intelligence
0.90
regional accents
0.80
chinese dialects
0.80
voice model
0.70
global benchmark
0.60
us rivals
0.50
openai
0.40
xai
0.40
speech arena leaderboard
0.40
§ 07
Topic connections
Interactive graph Network visualization showing 26 related topics
View Full Graph Person Organization Location Event|Click node to navigate|Edge numbers = shared articles