Open Source AI

Nous Research reportedly raised $90 million for Hermes, an open-source AI agent that could expand competition in assistants and enterprise AI.

Mistral has launched Mistral Large 4 as an open-weight model, giving AI teams another model to assess for controlled deployment and customization.

Cloudflare has introduced Clef, open-source decision models and an RL fine-tuning platform, signaling a push toward trainable AI control systems for builders.

AWS Strands Labs has released the open-source Strands Decider 2B, a 2-billion-parameter model aimed at faster, lower-cost option selection.

A Center for Data Innovation commentary urges Europe to strengthen its open-source AI strategy, but offers no new policy or product details.

NVIDIA has released OpenShell 0.1.0, an open-source runtime that adds enforceable permissions, sandboxing, and credential controls to AI agents.

Supersonic Labs has released Julia 1, a 144.3M-parameter open decision model designed to run on CPUs, widening options for local AI deployment.

IBM and NASA reportedly released an open-source Lunar Foundation Model and SomBench Dataset, putting new tools and a reported 23% error cut in reach.

Ant Group has released Ling-3.0-flash-Fin as an open-source finance model, giving developers a new option for research and financial workflow tools.

Reports link MiniMax M3 to a Saudi national-level client, highlighting how Chinese open-source AI models may be entering strategic overseas deployments.

MBZUAI’s Institute of Foundation Models says K2 Horizon opens weights, code, data and methods, challenging closed AI on reproducibility.

A Help Net Security report spotlights Halo-record, an open-source effort aimed at making AI agent activity easier to inspect, trace, and govern.

Meta has reportedly open-sourced its most powerful AI model to challenge OpenAI and Anthropic, putting its escalating AI spending under investor scrutiny.

Block’s open-source Goose gives developers a local alternative to Anthropic’s Claude Code, sharpening debate over AI coding costs, limits, and privacy.

A Doug Levin analysis argues that China’s open-source AI strategy is reshaping model competition, but the available evidence does not verify its claims.

Fortune reports that Chinese open-source AI is attracting U.S. business interest, raising new questions about cost, performance, trust, and deployment.

Z.ai confirmed it built Ox Alpha, an anonymous open-weight model leading early benchmarks, and plans to release its weights for coding and agentic workloads.

Alibaba’s Qwen models reportedly passed 3 billion downloads, raising questions about how open-model adoption is measured against Meta and Google.

Zhipu AI has released GLM-5.3, touting stronger open-weights coding and cybersecurity performance while postponing its weights release for reviews.

Block’s open-source Goose offers local AI coding agents without subscriptions, challenging Claude Code’s premium pricing while exposing trade-offs in quality and hardware.