Releases

Google Cloud announces Ironwood, its seventh-generation TPU designed for large-scale inference and thinking models.

Anthropic launched Claude Max with substantially higher usage limits for frequent individual users.

IBM introduced z17 with Telum II and support for accelerated enterprise inference.

AWS released a unified speech-to-speech foundation model for low-latency conversational applications.

Runway released the faster and cost-efficient Gen-4 Turbo video model.

Meta released a multimodal safety classifier for Llama 4 inputs and outputs.

Meta previewed Behemoth as a teacher model for the Llama 4 family; it was not made generally available.

Meta released Llama 4 Maverick, a natively multimodal mixture-of-experts model for general AI workloads.

Meta released Llama 4 Scout, a natively multimodal mixture-of-experts model with a 10-million-token context window.

Genspark said its first-generation AI Search product had reached five million users, then shifted focus to the more general Super Agent.

Hugging Face reported that Gradio had reached one million monthly users, reflecting its adoption for open AI demos and applications.

Microsoft updates Copilot with personalization through Memory, task execution through Actions, Deep Research and expanded Vision capabilities.

Runway raised $308 million in a General Atlantic-led round at a valuation above $3 billion.

V7 introduced Omni Reference for carrying subjects and visual elements into new images.

V7 introduced a fast, lower-cost Draft Mode for rapid iteration.

Midjourney released V7 with improved prompt precision, image coherence, Draft Mode, and Omni Reference.

xAI made Grok 3 flagship models generally available through its API.

Anthropic announced its first developer conference focused on building with Claude and Claude Code.

Genspark launched Super Agent, an autonomous assistant that plans, acts, and uses tools across everyday tasks.

CoreWeave submitted the first cloud-provider MLPerf Inference v5.0 results for NVIDIA GB200, reporting more than 800 tokens per second on Llama 3.1 405B.

Anthropic launched Claude for Education for universities, including a learning mode and institution-wide deployments.

Moonshot published the reasoning-oriented Kimi-VL A3B Thinking checkpoint.

Moonshot published the instruction-tuned Kimi-VL A3B checkpoint.

Moonshot introduced Kimi-VL, an open-weight mixture-of-experts vision-language family.

Baidu officially released version 3.0 of its deep-learning framework.