Alibaba Unveils Qwen3.8-Omni-Flash, Narrowing the Gap with Gemini 3.8 Flash in Multiple Performance Metrics
20 hour ago / Read about 0 minute
Author:小编   

Alibaba's Qwen team has recently rolled out a cutting-edge native full-modality model, Qwen3.8-Omni-Flash. This innovative model seamlessly blends audio and video comprehension capabilities with AI agent functionalities, accommodating four distinct input types: text, image, audio, and video. Moreover, it boasts an impressive context window of 1 million Tokens. As per Qwen's official announcement, when pitted against its predecessor, Qwen3.5-Omni-Plus, the Qwen3.8-Omni-Flash model has demonstrated an average performance enhancement exceeding 25% across 29 benchmark assessments. Concurrently, it has substantially slashed the expenses associated with audio and audiovisual processing.