Qwen3.8-Omni-Flash: an omnimodal model with one million tokens, but without open weights
Alibaba publishes a model that thinks natively about audio and video and knows how to call tools. Unlike previous Qwen, it remains reserved for the API.
AM-IA brings together news, research papers and applications relating to artificial intelligence published around the world into a single feed — and makes them available in your European language at the click of a button.
Scroll to browse · Click to filter
Alibaba publishes a model that thinks natively about audio and video and knows how to call tools. Unlike previous Qwen, it remains reserved for the API.
The Tokyo-based laboratory relies on a learned orchestrator rather than a single model: Fugu routes tasks to a pool of models and claims better scores for less.
The project aims to help code agents identify, reproduce and fix vulnerabilities.
Without changing its foundation, SpaceXAI pushes Grok to a level close to GPT-5.6 Sol thanks to additional training targeted on agentic tasks.
Comparison of the main development agents assisted by IA, their modes of integration and the uses to which everyone best responds.
The largest model ever produced by Alibaba is based on a mix of experts and only activates 95 billion parameters per token generated.