Alibaba Qoder Voice Tested: Can It Finally Ditch the ‘Li Jie’ Meme?
The author evaluates Alibaba’s newly released Qoder Voice, a real‑time full‑duplex AI coding assistant powered by Qwen‑Audio‑3.0‑Realtime, describing its setup, interaction flow, model performance (84.1% score, global No.1), and practical experience including interrupt handling, Appshot context, and shortcut configuration.
Alibaba’s Qoder, a popular domestic coding tool, has added a new feature called Qoder Voice , which the author describes as the first real‑time voice‑interactive programming agent in China.
Qoder Voice uses the real‑time full‑duplex speech model Qwen‑Audio‑3.0‑Realtime , which scores 84.1% on the Artificial Analysis benchmark, ranking #1 globally. Unlike previous one‑question‑one‑answer voice interfaces, this model allows users to interrupt, ask follow‑up questions, or adjust requirements at any time, with the system responding instantly.
The voice capability is launched from the Qoder Quest mode by clicking Start new voice chat . Once a voice command is issued, Qoder automatically creates a task, invokes tools, and continues executing the task in the background while maintaining the conversation.
A floating voice window can be summoned from any interface, and the author recommends assigning a shortcut key (found under Settings‑Voice‑Real‑time Voice) to toggle the assistant. The shortcut also supports adjustments for voice tone and speed.
In addition to voice input, Qoder offers Appshot , a feature similar to Codex’s screen‑snapshot capability that captures the active window’s image and text, providing contextual information for more accurate responses.
During personal testing, the author notes that Qoder Voice keeps up with rapid speech, automatically corrects recognition errors, and allows continuous dialogue without waiting for a page to finish processing. The author recorded a video demonstration showing the system handling fast speech and on‑the‑fly corrections.
Two standout aspects are highlighted:
The tool bundles the fastest, most comprehensive, and strongest large models, including the “Extreme Mode” and the free‑trial Qwen‑3.8‑Max‑Preview , which offers ample credit for extended use.
The “expert team” mode proves highly effective; the author’s own article about visualizing scientific concepts was broken down into 10 interactive experiments, creating nine new development scenarios, documenting them, and deploying the solution, all driven by more than 20 expert agents. Historically, domestic AI coding tools were perceived as functional but lacking polish compared to Claude Code, Codex, or Cursor. The author now feels that Qoder has reached a level of usability that makes it a strong contender. Qoder Voice is currently available on the international version of Qoder (download at qoder.com/download). Interested users can try it and assess its conversational efficiency.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Old Zhang's AI Learning
AI practitioner specializing in large-model evaluation and on-premise deployment, agents, AI programming, Vibe Coding, general AI, and broader tech trends, with daily original technical articles.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
