All AI News

Browse AI developments, their sources and dates.

1281 updatesLatest update 2026-10-05 09:10 UTC+8
Clear filters

Latest updates

1281 updates · UTC+8

2026-09-19

6 updates

2026-09-18

16 updates
千问 QwenModels

Qwen3.8-LiveTranslate: Know the Person, Convey the Meaning

Simultaneous interpretation is not just about translating quickly; it is even more about hearing clearly and translating accurately. Qwen3.8-LiveTranslate reconstructs real-time simultaneous interpretation with the Interleave architecture, achieving comprehensive improvements in faithfulness, fluency, and conciseness, with latency per average audio length (LAAL) reduced from 2.8 seconds to 2.3 seconds. We hope that simultaneous interpretation not only conveys language, but also preserves information about the people and context in the communication. Therefore, on the basis of supporting 60 languages, Qwen3.8-LiveTranslate adds three new capabilities to make simultaneous interpretation more broadly applicable in real-world scenarios: real-time speaker separation, so that each sentence is clearly attributed and voice cloning is more stable; source text and translation output in the same frame, with both languages on the same screen; and long-context disambiguation, so that the present is understood in connection with the preceding text and the translation of names and terminology is more accurate.

Qwen3.8-LiveTranslate Features Overview
Read here Read the source
千问 QwenModels

Qwen3.8-Omni-Flash: Sharp Ears, Keen Eyes, and Highly Capable

Today, the new-generation natively omni-modal model Qwen3.8-Omni-Flash officially goes live. The core goal of this model is to enhance its Agent capabilities in real-world productivity scenarios, pushing omni-modal models from "understanding omni-modal content" further toward "planning tasks, calling tools, and completing creation." Building on general Agentic capabilities such as coding, text-based knowledge work, and GUI operation, Qwen3.8-Omni-Flash further expands Agentic applications centered on audio and video, achieving remarkable results in workflows that require integrated processing of text, images, audio, and video, such as video editing, music video creation, film production and narration, audio-video-to-text-and-image summarization, and audio-video conversation.

Read here Read the source