Qwen3.5-Omni learned to write code from spoken instructions and video without anyone training it to

2026-03-31 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, quick

Summary

Alibaba has introduced Qwen3.5-Omni, an advanced omnimodal AI model capable of processing and understanding text, images, audio, and video inputs. This new model demonstrates strong performance, reportedly surpassing Google's Gemini 3.1 Pro in audio-related tasks. A notable, unexpected capability discovered during its development is its ability to generate code based on spoken instructions combined with video input, showcasing a novel multimodal interaction for programming. The release highlights Alibaba's continued advancements in developing comprehensive AI systems that integrate various data types.

Key takeaway

For research scientists evaluating multimodal AI models, Qwen3.5-Omni's reported audio performance and emergent code generation from spoken and video input warrant close examination. You should consider benchmarking its capabilities against existing models like Gemini 3.1 Pro, particularly for tasks involving complex audio understanding or novel programming interfaces.

Key insights

Qwen3.5-Omni is an omnimodal AI model excelling in audio and unexpected code generation from multimodal input.

Principles

• Multimodal integration enhances AI capabilities.
• Unexpected emergent behaviors can arise in complex models.

In practice

• Use Qwen3.5-Omni for advanced audio processing.
• Explore multimodal code generation from speech and video.

Topics

• Qwen3.5-Omni
• Omnimodal AI
• Code Generation
• Multimodal AI
• Alibaba

Best for: Research Scientist, AI Scientist, Machine Learning Engineer, AI Engineer

Related on AIssential

• Qwen3.5-Omni is here! Scaling up to a Native Omni-modal AGI — Alibaba's Qwen3.5-Omni unifies diverse modalities and advanced conversational features into a single, highly capable AI system.
• Qwen 3.5 The GREATEST Opensource AI Model That Beats Opus 4.5 and Gemini 3? (Fully Tested) — Qwen 3.5 is a powerful open-source, multimodal AI model excelling in speed and diverse applications.
• Nemotron 3 Nano Omni Local Test | Document Understanding, Audio Processing, Coding, Audio | 🔴 Live — Nemotron 3 Nano Omni is a multimodal MoE model from Nvidia, showing promise in audio but struggling with complex document understanding.
• 9 demos of Gemini Omni and Gemini 3.5 in action — Gemini Omni and 3.5 Flash advance multimodal AI, enabling conversational video editing and intelligent agentic workflows.
• Qwen 3.5 Local Test with Ollama | Coding, OCR, Data Extraction, Image Understanding — Qwen 3.5 MoE model offers strong multimodal capabilities and efficient local inference.
• 🗞️ Alibaba released their open-source Qwen3.5-9B model, and it holds its own against OpenAI's 120B param gpt-oss system. — Smaller, efficient AI models are achieving performance comparable to much larger systems through architectural innovations and multimodal training.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.