60x faster cold start: treating peer GPUs as weight servers
The Runway platform team developed the NCCLBack system, which shortens model cold-start time from minutes to seconds…
AI news monthly archive: May 2026 features 678 curated AI news items covering model releases, product updates, industry news, papers, tutorials and opinions, each with an AI summary, source credit and original link.
The Runway platform team developed the NCCLBack system, which shortens model cold-start time from minutes to seconds…
Goodfire Research published a study showing that models spontaneously express awareness of being evaluated in their…
Ad-supported AI based on open-source models and commercial GPUs is economically viable. Calculations show that a…
A gallery for sharing codex pets: submissions are now open, submit via the link below 👇
Introducing response caching: save significant money and time on testing and agent retries. Blog post:…
New feature: the "-latest" model alias 🔀 routes requests to "~anthropic/claude-opus-latest", "~openai/gpt-latest" and…
In a fireside chat at Sequoia Capital's AI Ascent, Demis Hassabis demonstrated his exceptional capacity for…
The US Department of Defense has reached agreements with seven to eight leading AI companies to integrate their AI…
The US Pentagon announced agreements with eight leading AI companies — SpaceX, OpenAI, Google, NVIDIA, Reflection,…
You can now log in to OpenClaw with your ChatGPT account and use your subscription there! Happy lobster catching.
Voice cloning is now available through the xAI API! Create a custom voice in under 2 minutes, or choose from our…
Reinforcement learning can improve model performance in known domains, but in unknown domains it may cause models to…
In an interview, DeepMind founder Demis Hassabis looked back on his path from chess prodigy to the Nobel Prize in…
Bring your workflows into Codex in just a few clicks. Import settings, plugins, agents, project configurations and…
OpenRouter has officially launched text-to-speech and audio transcription. Through two new API endpoints, the platform…
Code with Claude, our developer conference, returns next week. Whether you are just getting started with Claude Code…
One week after launch, GPT-5.5 is already our strongest model release to date. API revenue is growing more than twice…
τ-voice is a new benchmark for real-time voice agents, containing 278 grounded customer service tasks across retail,…
This update improves model gateway integration, listing available models directly in the model selector when a…
Several recent developments show AI's substantial positive impact. In healthcare, Mayo Clinic's AI can detect…
This paper proposes a new method that embeds evaluation into the real-time execution loop of tool-calling agents. By…
@xai's new Grok-4.3 model is now live on OpenRouter! Grok-4.3 launches at a lower price than Grok-4.2 while delivering…
Midjourney V8.1 is now available on Discord and the official website. The update focuses on sharper, higher-quality…
Find a look for your website. Explore every direction at once. Define the goal, set the aesthetic, then let Luma…
A specific image-generation prompt for GPT Image 2 is going viral on social media. The prompt's core instruction is to…
Alphabet shares rose about 9%, adding roughly $420 billion in market value in a single day — the largest one-day gain…
Using Codex for daily work has never been easier. Choose your role, connect the apps you use every day, and try…
How do people seek guidance from Claude? We analyzed 1 million conversations to understand what people ask, how Claude…
A new mechanism called "auto-review" offers a safer default for deploying code agents. Through an independent reviewer…
The newly launched Response Caching header implements a caching mechanism for API requests: identical requests receive…
OpenRouter has added a Response Caching header to cache identical API requests, significantly reducing response times…
The AntLingAGI team announced the official open-source release of the Ling-2.6-1T model, now available on Hugging…
The Notebooks feature in the Gemini App is now officially available to free and paid users on mobile, with plans to…
Last week Google officially released Gemini Embedding 2, its first native multimodal embedding model, to the public.…
GitHub Copilot CLI offers two main usage modes: interactive and non-interactive. Interactive mode lets users build and…
Google's Gemini app has officially rolled out the Notebooks feature to free and paid mobile users, with expansion to…
Kostiantyn Vlasenko, a project manager with no programming experience, used Claude Code to build the stress management…
ARC Prize analyzed 160 reasoning replays of GPT-5.5 and Opus 4.7 on ARC-AGI-3; the two models scored 0.43% and 0.18%…
Page 17 of 17
Sponsored
Orange Relay (pro.nat6.net) is a premium AI API relay — an excellent transfer station with competitive pricing, reliable quality and broad model coverage including Claude, GPT and Gemini.
Visit Orange Relay