Alibaba's Qwen Audio 3.0 TTS Plus tops the competition in the text-to-speech rankings
Alibaba's Qwen Audio 3.0 TTS Plus has taken the top spot on the Artificial Analysis Speech Arena leaderboard, offering high quality output across 16 languages.
- Qwen Audio 3.0 TTS Plus ranks first on the Artificial Analysis Speech Arena.
- The model supports 16 languages and allows style control via tags or text.
- Generation speed is 16 characters per second, slower than top competitors.
Alibaba has released Qwen Audio 3.0 TTS Plus, a text-to-speech model that currently ranks first on the Artificial Analysis Speech Arena. This benchmark evaluates audio quality through blind testing, placing the new model ahead of competitors like Sonic and Simba in terms of output fidelity.
The model supports 16 languages and allows users to manipulate the speaking style using natural language prompts or specific tags such as [angry]. This provides fine-grained control over the emotional tone of the generated audio, which is useful for dynamic content creation.
Despite its high audio quality, the model lags in generation speed. It produces audio at roughly 16 characters per second, which is significantly slower than leading rivals Sonic 3.5 and Simba 3.2.
Provides access to a high quality, multilingual TTS model with granular style controls.
Enables better localization and voice agent capabilities with realistic audio.
Marks a step forward in AI voice realism and emotional expressiveness.
- Speech Arena
- A blind evaluation benchmark by Artificial Analysis that compares text-to-speech models based on user preference.
AI ToolsSnowflake Cortex, Explained Like an AI That Lives Next to Your Data
AI ToolsStop Letting AI Write Security Bugs: Introducing "hallint"
AI ToolsThe smolagents sandbox broke 'a, *b = list', one of Python's most common lines
AI ToolsAI’s most important protocol is getting a little bit easier to use
AI ToolsAI agents, meet the Azure Cosmos DB vNext emulator
Researchers gather at Cornell Tech to explore AI for molecular science - Cornell Chronicle
Researchers from Cornell Tech gathered to discuss the potential of AI in molecular science, exploring its applications and advancements.
SecurityUS threatens sanctions against Chinese AI models over IP theft
Treasury Secretary Scott Bessent warned the United States could impose sanctions on Chinese AI models if they are found to have stolen intellectual property, marking a tougher stance on China's AI progress.
HardwareNVIDIA Vera Rubin Driving Performance Per Watt, Lowest Token Cost for Partners Worldwide
NVIDIA has launched its Vera Rubin NVL72 platform, designed for gigascale AI, with production ramping up at major cloud partners including CoreWeave, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure. The platform aims to deliver high performance per watt and low token cost for AI workloads.
AI ResearchIntroducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
DeepMind has introduced Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber, new models in their Gemini series.
HardwareBuilt for Vera Rubin, NVIDIA Spectrum-6 Arrives in Gigascale AI Factories
NVIDIA has released Spectrum-6, a new Ethernet networking platform designed to connect hundreds of thousands of GPUs in gigascale AI factories.
Health systems need to prepare for an AI-driven reinvention of work, Abundant Alliance's Eric Langshur says - Fierce Healthcare
Abundant Alliance's Eric Langshur warns healthcare systems to prepare for an AI-driven reinvention of work.