A live feed of notable AI product, release, pricing, research, and developer updates gathered from monitored public sources.
Every Gradio Space on Hugging Face now automatically exposes a machine-readable /agents.md endpoint, allowing AI coding agents to discover and interact with spaces without manual configuration.
DeepSeek-V4, the latest model from Chinese AI lab DeepSeek with a one-million-token context window, has surfaced in the Hot articles section of the Hugging Face community blog.
StepFun has introduced StepClaw, a new conversational AI assistant on its platform offering web search, knowledge base Q&A, and image generation capabilities.
Baseten has updated its model inference platform with a restructured three-tier pricing model — Basic ($0/month pay-as-you-go), Pro, and Enterprise — adding HIPAA compliance, priority GPU access, and self-host deployment options.
Baseten has announced the deprecation of two model APIs while simultaneously launching discounted cache token pricing for all remaining model endpoints.
Fireworks AI has entered the model training space with a new platform preview that lets teams train and deploy frontier-scale models on a single unified system.
fal.ai has a live, publicly accessible pricing page for its fal-ai/flux/dev model endpoint, though no substantive pricing data was captured in this source retrieval.
fal.ai has updated its serverless platform dashboard with a new feature that displays why a runner was terminated directly alongside the runner's state.
The llama.cpp project shipped a quick bug-fix release on April 24, 2026, addressing a critical flaw in its structured output parser.
DeepInfra has added GLM-5.1, described as a state-of-the-art agentic engineering model, to its cloud inference catalogue.