Chinese and community-driven open-weight models like Kimi K3, DeepSeek, and Qwen are rivalling top proprietary LLMs, accelerating global competition and democratisation.
The pace of open-source releases, with models like Minimax 3 Pro at 2.7 trillion parameters and GLM 5.3, marks a turning point. As enterprise trust in closed vendors wanes — forced to “distill” client knowledge to justify trillion-dollar valuations —...
Moonshot’s open-source Kimi K3 sent tech and chip stocks tumbling, reviving memories of the DeepSeek shock. Yet the selloff reveals more about unsustainable cloud infrastructure spending than about the model’s capabilities, pointing to a structural r...
Axios reports that the Chinese open-source model Kimi K3 matches Anthropic’s Claude Opus, erasing the US lead. A turning point for on-premise deployment and data sovereignty.
Z.ai, the Chinese startup behind GLM, projects $1bn in annual sales while giving away its most powerful LLMs free. A model that challenges paywalled APIs and shifts value toward concrete deployment, with direct implications for those choosing on-prem...
Moonshot's free Kimi K3 model has reportedly matched the performance of Anthropic's Opus, accelerating the convergence between open and paid AI. For organizations evaluating on-premise deployment, this signals that proprietary-grade quality is becomi...
The new Kimi K3 model storms arena.ai, beating GPT-5.6 and Claude Fable, previously labeled 'too dangerous'. But how much do these benchmarks matter for on-premise deployment decisions? A different approach is needed: testing on real data and local h...
The Chinese Moonshot AI model climbs the independent ranking and overtakes Anthropic. Less mainstream noise, more substance for those eyeing high-performance, cost-efficient self-hosted LLMs.
Moonshot is set to release Kimi K3, the largest open Chinese LLM with an estimated 2 to 3 trillion parameters, aiming to rival Anthropic’s Opus 4.8. Its sheer scale raises tough questions for on-premise deployment: aggressive quantization, VRAM costs...
Thinking Machines has released Inkling, its first open-weight model. This move aligns with the trend of permissively licensed LLMs, designed for companies that want to run inference on their own infrastructure. We analyze what it means for data sover...
The Chinese lab is gearing up for a 2027 IPO after its first private raise. The chase for a record valuation could shift the open-weight balance and push self-hosting enterprises to reassess single-vendor model dependency risks.
The Trump administration and industry groups are discussing a streamlining of US open model releases, with a bound: equal or lesser capability to leading Chinese open models. A move that reshapes the open source AI landscape and on-premise infrastruc...
Hugging Face CEO Clem Delangue sees enterprises increasingly turning to open models for cost, accessibility, and data ownership. If production AI runs on open source, frontier models risk becoming less relevant, with deep consequences for hardware, l...
A Reddit thread highlights a structural truth: Chinese open-source labs lead benchmarks while American counterparts lag. The reasons—hardware restrictions, divergent incentives, and industrial strategy—are reshaping the field. For those considering o...
Zhipu founder Tang Jie argues that frontier AI safety hinges on openness and collective oversight, clashing with Chinese state control. A friction that redefines the boundaries of digital sovereignty.
