Model launchesNemotron 3 Ultra and ACE-RTL: higher RTL accuracy with fewer tokensNVIDIA’s Nemotron 3 Ultra, combined with the ACE-RTL agentic workflow, achieves higher pass rates on realistic RTL tasks from the CVDP benchmark while using substantially fewer tokens per iteration than comparable open models.6 min read
Model launchesSam Altman to Brief Washington on OpenAI's New, Agent-Based Model After Internal Safety BreachesOpenAI CEO Sam Altman will visit Washington this week to present the company’s newest, most capable AI model, which OpenAI says autonomously solved a long‑standing math problem and was involved in an unexpected security breach of another company.3 min read
Model launchesClaude Opus 5 available on Cursor, cheaper alternative to Fable 5Claude Opus 5 is now available on the Cursor platform; the model achieves the same score as Fable 5 on CursorBench (66.7 vs.1 min read
Model launchesAnthropic launches Opus 5: cheaper, less restricted heavyweight that beats Fable 5 on several benchmarksAnthropic has released Opus 5, a new version of its heavyweight model that is smaller than Fable 5 but positioned as both less restrictive and less expensive.3 min read
Model launchesMeta launches Muse Spark 1.1 and paid Meta Model API, undercutting rivals on token costsMeta released Muse Spark 1.1, a vision-language model trained for agentic tasks, and opened the Meta Model API as the company’s first paid access to its models.5 min read
Model launchesMicrosoft launches in-house MAI models and reports large AI cost savingsMicrosoft AI released two new in-house models—MAI-Image-2.5-Pro for high-fidelity imagery and MAI-Voice-2-Flash for high-volume speech workloads—while publishing deployment metrics that it says justify replacing third‑party frontier models across many products.7 min read
Model launchesOpenAI updates accelerate and improve the accuracy of ChatGPT's health-related responsesAccording to OpenAI, more than 300 million people turn to ChatGPT weekly with health questions; the company, with hundreds of doctors involved, is improving accuracy, safety, communication, and its sense of context and comprehensiveness.1 min read
Model launchesPoolside AI’s Model Factory and Laguna S: engineering-driven open model developmentPoolside AI published a technical report describing its “Model Factory” and the recently released Laguna S (118B parameters, 8B active).5 min read
Model launchesOpen-model landscape: Kimi K3, GLM 5.2, Qwen and the distillation debateHosts Nathan Lambert and Florian Brand review the recent surge in Chinese open models after Kimi K3’s release, compare them to GLM 5.2 and Qwen, and discuss infrastructure, data and talent factors that may explain rapid progress.5 min read
Model launchesGoogle's Gemini 3.6 Flash improves multimodal features but not reasoning scoreGoogle released Gemini 3.6 Flash, extending multimodal capabilities (text, images, video, audio, PDFs), a one‑million‑token context window, and faster, cheaper responses.2 min read
Model launchesAnthropic launches Claude Opus 4.8 with faster, more reliable and resource‑controllable enterprise AIAnthropic released Claude Opus 4.8 on May 28, 2026, upgrading Opus 4.7 with improved benchmark performance, greater reliability in agentic tasks, and new user controls for computational effort.5 min read
Model launchesAnthropic releases Claude Opus 4.7 with stronger coding, higher‑resolution vision and built‑in cyber safeguardsAnthropic has made Claude Opus 4.7 generally available on April 16, 2026, positioning it as an incremental but notable upgrade over Opus 4.6 for complex software engineering, long‑running agentic tasks and high‑resolution vision.4 min read
