Ant Group's Ming-3.0-tts Tops Artificial Analysis Voice Arena
10/19/2026 — 10/20, 11:19·3 sources (all)·4 reports (all)
Key takeaways
- Ant Group's Ming-3.0-tts has taken the top spot in Artificial Analysis's Controlled Voice TTS Arena, making it the highest-ranked open-weigh…
Background & analysis
On October 19, 2026, inclusionAI released Ming-3.0-audio-flash, a native audio-language model, on Hugging Face. The model uses a Mixture-of-Experts architecture with 126B total parameters and 6.5B active parameters, and is available under the MIT license. According to its model card, it accepts audio together with text instructions and supports transcription, spoken question answering, long-recording understanding, and reasoning about speech, environmental sounds, and music. It can also return an optional spoken response. Chinese and English are listed as its primary languages, with claimed support for accents, dialects, and noisy recordings.
AI-generated from 3 reports · updated 1 hour ago
Latest turnArtificial Analysis says Ant Group's Ming-3.0-tts is now the top open-weights model on its Controlled Voice TTS Arena and ranks first on its Pronunciation Robustness leaderboard, level with Fish Audio's Drama 3 Preview and ahead of Eleven v4. Both placements come from the third-party evaluator's leaderboards rather than the model's developer.
- Heat index
- 261
- All sources
- 3
- All reports
- 4
- First seen
- 13 hours ago
Reports on this story headlines open the original
Artificial Analysis says Ant Group's Ming-3.0-tts is now the top open-weights model on its Controlled Voice TTS Arena and ranks first on its Pronunciation Robustness leaderboard, level with Fish Audio's Drama 3 Preview and ahead of Eleven v4. Both placements come from the third-party evaluator's leaderboards rather than the model's developer.
X：Artificial Analysis (@artificialanlys)First-partyAI score 51Ant Group's Bailing team has open-sourced Ming-3.0, a family of five speech models covering text-to-speech, audio understanding and dialogue, and speech recognition, which developers can use separately or combine in a workflow. The TTS models accept text descriptions of a target voice plus a director mode for emotion shifts, stressed words and pacing, while the audio models transcribe and summarize long recordings while retaining emotion and speaker information.
IT之家AI score 63
inclusionAI has published Ming-3.0-audio-flash on Hugging Face, a 126B-parameter MoE with 6.5B active parameters that takes audio plus text instructions and returns text or a spoken response, released under MIT. The model card lists transcription, spoken Q&A, long-recording understanding and reasoning about speech, environmental sounds and music, with Chinese and English as primary languages and claimed support for accents, dialects and noisy recordings.
Reddit · LocalLLaMAAI score 64inclusionAI has released Ming-3.0-audio-flash on Hugging Face, a native audio-language Mixture-of-Experts model with 126B parameters. It takes audio alongside text instructions and supports speech recognition, spoken question answering, audio understanding and reasoning, with optional synthesized-speech responses.
Reddit · LocalLLaMAAI score 54
Other stories people are talking about
- Heat index 88NewArtificial Analysis launches HealthBench leaderboards1 sources
- Heat index 77Artificial Analysis launches AA-PersonalityBench to score model personality1 sources
- Heat index 1126RisingMiniMax M3.1 Flash launches on Cline and gmi_cloud15 sources
- Heat index 898Anthropic pledges 5% stake to public benefit fund, with top-ups as valuation grows16 sources
- Heat index 519SurgeNscale in talks with Anthropic over roughly $45B compute deal6 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion