Cohere released Command A+, a sparse mixture-of-experts open-source model designed for agentic tasks. The model features 218 billion total parameters with 25 billion active parameters and is licensed under Apache 2.0.
Cohere has released Command A+, marking its first open-source model under the permissive Apache 2.0 license. The sparse mixture-of-experts (MoE) architecture contains 218 billion total parameters but activates only 25 billion during inference, optimizing computational efficiency.
The model is purpose-built for agentic tasks—applications where AI systems operate autonomously to complete multi-step objectives. This positioning reflects the growing industry focus on AI agents as a critical frontier beyond traditional language model applications.
The sparse MoE design is significant. Rather than activating all parameters for each inference, the architecture selectively routes inputs through relevant expert networks, reducing computational overhead while maintaining performance capabilities. This approach enables deployment on more modest hardware compared to dense models of equivalent scale.
Cohere's move to open-source under Apache 2.0 represents a strategic shift toward broader accessibility. Unlike some restrictive licenses, Apache 2.0 permits commercial use and modification, allowing developers and organizations to integrate Command A+ into proprietary applications without licensing complications.
The release follows Cohere's announced merger with German AI startup Aleph Alpha, signaling the company's broader expansion in the competitive AI landscape. The open-source release positions Cohere alongside other players developing foundation models for enterprise and research applications.
Command A+ joins a growing category of open agentic models as organizations increasingly pursue autonomous AI systems. The combination of open licensing, efficient architecture, and agent-focused design makes the model relevant for developers building on-premises or private cloud deployments where proprietary model dependencies pose constraints.
Cohere's focus on sparse architectures and agent capabilities reflects market demands for efficient, practical AI systems beyond general-purpose chat interfaces. As agentic AI gains traction, model efficiency becomes increasingly important for real-world deployment at scale.
Z.ai released GLM-5.3's weights on Hugging Face under a new license that requires large companies to undergo security review before hosting the model. The change marks a departure from the standard MIT license.
Anthropic has introduced the Model Hardware Standard (MHS), a unified interface enabling AI agents to operate robotic arms, lab instruments, and other physical devices. Early testing shows integration time has dropped from weeks to hours.
Open-weight AI companies—those releasing freely available models—are attracting major acquisition interest from tech giants. The trend reflects growing capital investment in the business model of distributing AI models at no cost.
Google Deepmind has upgraded its Co-Scientist AI system to autonomously plan experiments, operate lab equipment, and publish scientific papers. The Gemini-based multi-agent platform demonstrated experimentally validated results across materials science, chemistry, and medical AI development.