Models like GLM‑5.2 and Mistral lead a wave of open‑weight releases, sparking community innovation and debates around permissive licensing for local use.
After firing AI chief Junyang Lin, Qwen has kept its 3.7 model line fully closed source, making it the only major Chinese lab without a recent open release. As competitors like DeepSeek and GLM continue shipping, the move narrows choices for on-premi...
Noema AI has released Noema Atlas, an open-source peer-to-peer software for LLM weight distribution. Built on Iroh and BLAKE3-based content-addressing, it enables verified machine-to-machine transfers, automatic deduplication, and optional fallback t...
When the number two contender celebrates the top open source model, the message goes beyond niceties. z.AI's praise sparks a reflection on open LLM maturity, self-hosting strategies, and how such models reshape data sovereignty and TCO for enterprise...
The new AA-Briefcase benchmark evaluates LLMs on agentic knowledge work. Chinese model GLM-5.2 outperformed GPT-5.5, highlighting how specialized evaluations are reshaping model selection—also for self-hosted deployments where control and data sovere...
GLM-5.2 has been recognized as the top "open weight" Large Language Model (LLM) for creative writing, according to Sam Paech's benchmark on EQ Bench. This achievement highlights the potential of accessible models for on-premise deployment scenarios, ...
Z.ai has open-sourced its GLM 5.2 model, generating significant community excitement. Developers and enterprises are now eagerly anticipating a "Flash" series successor, ideally within the 27 to 120 billion parameter range, to optimize on-premise and...
The GLM-5.2 model has achieved the first position in the Design Arena ranking, surpassing Claude Fable 5, which is now unavailable. This result highlights the dynamic nature of the Large Language Models landscape and raises questions about the stabil...
Intense speculation surrounds "Le Gros Chaton," a rumored new model from Mistral AI. It's whispered to possess exceptional capabilities, including a one-billion-token context window, potentially surpassing current market leaders. The crucial question...
A new open-weights LLM, Qwable-v1, has been released, derived from Anthropic's controversial Claude Fable-5. Distilled on a single H200 GPU, it offers agentic coding and tool-use capabilities, with GGUFs available for on-premise deployment, raising q...
A recent poll on X, conducted by z.ai, reveals declining support for Large Language Models with open weights distributed under an MIT license. With 1,800 votes cast and only a few hours remaining, the preliminary result suggests a potential shift in ...
