Qwen Audio Tops TTS Leaderboard, But Best Model Locked Behind API?
b111ue · reddit · 2026-07-22
Alibaba's Qwen-Audio-3.0-TTS-Plus recently took the #1 spot on Artificial Analysis's TTS leaderboard, beating out major models like Gemini and ElevenLabs. The samples sound genuinely impressive, natural, and highly controllable.
However, its distribution strategy has sparked discussion in the open-source community:
- Qwen3-TTS: Offers downloadable Apache 2.0 weights for local deployment.
- Qwen-Audio-3.0-TTS-Plus: Currently available only via Alibaba Cloud API, with no official weights released.
The author notes that while this doesn't mean Qwen is abandoning open source, their latest decisions show a slow shift toward commercialization with their strongest models. The community is now debating the trade-off: should developers prefer the best model locked behind an API, or a slightly weaker model whose weights they can download and modify locally?
Related event: Alibaba's Qwen TTS Tops Charts, API Access Only(3 posts)→
More from Models
- GPT-6.1 Terra spotted in the OpenAI API, teased by benchmark group LuminaBench — airesearch12 · 2026-10-27
- Ortet's first health eval: 34 open-weight models, 50 tasks, generic rankings fall short — thegautamkamath · 2026-10-27
- Fern 0.1: a 0.1–0.5T coding model one researcher pretrained in 3.5 months for under $10M — teortaxesTex · 2026-10-27
- VLM Run launches OCR Playground for side-by-side comparison of GLM-OCR, DeepSeek-OCR-2 and more — spillai · 2026-10-27
- GPT-6.1-Terra Live With Big Price Cut: $5 vs $12 Output Pricing — scaling01 · 2026-10-27
- Nat Lambert: traces from mid-tier open models are almost all super janky — natolambert · 2026-10-27