Model launchesAnthropic launches Claude Opus 4.6 with 1M-token context beta and upgraded coding and reasoningAnthropic announced Claude Opus 4.6 on Feb 5, 2026, a new Opus-class model with improved coding, longer agentic task performance, stronger long-context retrieval and a 1 million token context window in beta.5 min read
Model launchesAnthropic releases Claude Sonnet 4.6 with larger context window and improved computer‑use skills at same priceAnthropic has launched Claude Sonnet 4.6 (Feb 17, 2026), a Sonnet‑class model that improves coding, computer use, long‑context reasoning, agent planning, knowledge work and design while introducing a 1‑million‑token context window in beta.5 min read
Model launchesAnthropic launches Claude Opus 4.5: cheaper, more efficient coding, agents and long-context workflowsAnthropic released Claude Opus 4.5 on Nov 24, 2025, positioning it as an improved model for software engineering, agentic workflows and long-context tasks.4 min read
Model launchesGoogle launches token‑efficient Gemini Flash models for agentic and cybersecurity useGoogle DeepMind introduced three proprietary Gemini Flash models—Gemini 3.6 Flash, Gemini 3.5 Flash‑Lite, and the specialized Gemini 3.5 Flash Cyber—designed to reduce token usage and latency for agentic workloads.5 min read
Model launchesGoogle launches token-efficient Gemini Flash models to lower AI agent costsGoogle DeepMind introduced three proprietary Gemini Flash models—Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and the specialized Gemini 3.5 Flash Cyber—aimed at reducing token usage and latency for enterprise AI agents.4 min read
Model launchesOpenAI launches ChatGPT for small businesses with ChatGPT Work and GPT‑5.6OpenAI has introduced a ChatGPT for small businesses program to help entrepreneurs increase productivity and scale operations using AI.3 min read
Model launchesGoogle launches Gemini 3.6 Flash, 3.5 Flash‑Lite and a cyber‑focused 3.5 Flash variantGoogle is releasing three new models in its Flash family: Gemini 3.6 Flash, a more efficient and higher‑quality successor to 3.5 Flash; Gemini 3.5 Flash‑Lite, a low‑latency, high‑throughput model optimized for agentic workflows; and Gemini 3.5 Flash Cyber, a security‑focused model deployed via CodeMender to trusted partners.4 min read
Model launchesGoogle launches Gemini 3.6 Flash and 3.5 Flash‑Lite; limited 3.5 Flash Cyber pilot for CodeMenderGoogle introduced two new models in its Flash family — Gemini 3.6 Flash, focused on higher efficiency and quality, and Gemini 3.5 Flash‑Lite, aimed at low latency and high throughput for agentic workflows.4 min read
Model launchesMoonshot's open Kimi K3 model narrows US lead and raises questions about American AI strategyMoonshot's release of the open-source Kimi K3 model last week has prompted analysts to say it reduces Washington's edge in artificial intelligence, exposes limits of US chip export controls, and challenges the logic of massive AI spending by US tech firms.2 min read
Model launchesGoogle DeepMind unveils new Flash models focused on efficiency and a limited cybersecurity pilot; Gemini Pro update absentGoogle DeepMind introduced three new Flash family models—Gemini 3.6 Flash, 3.5 Flash‑Lite, and 3.5 Flash Cyber—aimed at improving efficiency, latency and cost for production AI agents.2 min read
Model launchesKimi K3 bemutatja az open-weight, hatékonyság-vezérelt versenykorszakot a nagymodell-ökoszisztémábanMoonshot AI július 16-án tette közzé Kimi K3-at, egy 2,8 trillió paraméteres MoE modellt, amelynek súlyait a cég ígérete szerint július 27-én teszik elérhetővé.4 min read
Model launchesKimi K3, a 2.8T open-weight Chinese model, nears Western frontier performanceKimi K3, a 2.8 trillion-parameter open-weight model from China, posts frontier-level results that often match or trail leading proprietary systems such as Claude Fable 5 and GPT 5.6 Sol.3 min read
