Models
Kimi K3 sets new benchmark with two million token context window
Kimi K3 has reportedly introduced a two million token context window, a significant advancement in A...
Category
Models
Releases, version updates, and benchmark results for frontier models like GPT, Claude, Gemini, and Llama. We go beyond press-release summaries to surface the metrics builders actually compare — pricing, context window, multimodal support, latency — and add Korean-market context where it matters. Sources include OpenAI, Anthropic, Google DeepMind, Meta AI blogs, arXiv, Hugging Face, and original reporting from The Information and TechCrunch.
Google DeepMind releases DiffusionGemma, an open model with four times faster local AI inference
Google DeepMind has introduced DiffusionGemma, a new open model in the Gemma 4 family designed for significantly faster local AI inference. Unlike tra...
Briefing Queue
Top picks right now
Category Digest
By category
Latest
Latest stories
Models
Apertus unveils open foundation model to support sovereign AI initiatives
Apertus has introduced an open foundation model specifically designed to facilitate sovereign AI eff...
Models
Google DeepMind releases DiffusionGemma, an open model with four times faster local AI inference
Google DeepMind has introduced DiffusionGemma, a new open model in the Gemma 4 family designed for s...