GO
| HSI1 | 24,788.14 | 0.00 | -- |
| HSCEI1 | 8,374.30 | 0.00 | -- |
| Back Zoom + Zoom - Block Traded | |
|
Alibaba Qwen Launches All-Modal Large Model Qwen3.5-Omni, Surpassing Gemini-3.1 Pro in Several Metrics
2026-03-31 07:16:19 According to domestic media reports, Alibaba Qwen has officially launched the latest generation all-modal large model Qwen3.5-Omni. This model achieved SOTA (State of the Art) results in 215 third-party performance tests, surpassing Google Gemini-3.1 Pro in several core metrics such as audio-video understanding and real-time interaction. On the technical level, the Qwen3.5-Omni series includes three sizes of Instruct versions: Plus, Flash, and Light, supporting a 256k long context. The model supports over 10 hours of audio input and more than 400 seconds of 720P (1FPS) audio-video input. It underwent native multi-modal pre-training on massive text, visual, and over 100 million hours of audio-video data, demonstrating exceptional all-modal perception and generation capabilities. Compared to Qwen3-Omni, the Qwen3.5-Omni has significantly enhanced multilingual capabilities, supporting speech recognition in 113 languages and dialects and speech generation in 36 languages and dialects. Currently, the Plus, Flash, and Light APIs of Qwen3.5-Omni are available on Alibaba Cloud's BaiLian platform. The API call pricing is less than RMB0.8 per million tokens, which is less than one-tenth of the price of Gemini-3.1 Pro. Alibaba stated that Qwen currently serves over one million customers across industries such as the internet, finance, consumer electronics, and automotive, maintaining the leading position in enterprise-level large model calls in mainland China. (ta/w)~ AASTOCKS Financial News URL: www.aastocks.com | |