Google DeepMind has officially launched its new frontier model, Gemini 4 Argon, achieving state-of-the-art results in 13 out of 19 major benchmarks. The model's standout feature is the experimental 'Long Decode Continuation' technique, which extends output capacity to 1 million tokens—an industry first. Currently, access is restricted to trusted testers in cybersecurity via the Fairwind Program, with broader availability promised soon.
State-of-the-Art
SOTA
Refers to the highest level of performance currently achieved in a specific task.
Curated Intelligence
OpenAI announced GPT-6 Astra, claiming it is the world's most intelligent and aligned model with state-of-the-art capabilities in computer use and software engineering. However, the text lacks specific performance metrics, benchmarks, or technical details, focusing instead on marketing vision and full-stack compute strategy.
OpenAI has officially released GPT-5.4, positioning it as an efficient frontier model for professional work. The update achieves state-of-the-art performance in coding, computer use, and tool search, while supporting a massive 1 million-token context window.
Zhipu AI and Huawei have open-sourced GLM-Image, a new multimodal model fully trained on Ascend Atlas 800T A2 chips using the MindSpore framework, validating domestic hardware for frontier AI. The model employs a hybrid 'autoregressive + diffusion decoder' architecture, achieving state-of-the-art performance in complex text rendering (especially Chinese characters) with ultra-low inference costs.