Open-source Chronicles-OCR benchmark evaluates visual perception of ancient Chinese characters
An open-source benchmark, Chronicles-OCR, has been released to evaluate the visual perception capabilities of vision…
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.
An open-source benchmark, Chronicles-OCR, has been released to evaluate the visual perception capabilities of vision…
Google and Blackstone Group announced a partnership to establish a new AI cloud services company, with Blackstone…
Claude AI assistant has released version v2.1.144. The update mainly adds `/resume` support for background sessions…
Hugging Face released six Ettin Reranker models (17m, 32m, 68m, 150m, 400m, 1b), based on the Ettin ModernBERT encoder…
Claude Platform introduced the MCP Tunnels research preview, which supports connecting to MCP servers within private…
We're on the road to I/O 2026. See you tomorrow at 10 a.m. Pacific Time!
You can now create more with Claude Design. We have doubled the token limits on all plans.
Three new models from @xai's Grok creative suite are now available on OpenRouter: • Grok Imagine Image Quality:…
In its verdict, the jury unanimously found that Elon Musk's lawsuit against OpenAI was filed beyond a reasonable time…
Prompt caching diagnostics are now available in the Claude Console. When a request misses the cache, you can now see…
GitHub has launched remote session control for Copilot and rolled it out fully. The feature lets users start a Copilot…
Musk tweeted asking for help improving Grok Build. User @morganlinton shared their experience using Grok Build to…
NVIDIA Cosmos Predict 2.5 is a 2B-parameter world model that generates physically plausible video from text, images or…
Google Research's Empirical Research Assistant (ERA) has now published a paper in Nature. It is a Gemini-based AI tool…
Google is consolidating its AI terminal tools, migrating the community-facing Gemini CLI to a new platform,…
The Google Tensor ML SDK has entered beta, enabling developers to build and deploy high-performance machine learning…
The Google AI Edge Gallery app expands on-device AI capabilities on Android by introducing experimental support for…
On May 22, xAI announced that users can now use the Grok model directly within the open-source personal assistant…
The Google Cloud and NVIDIA developer community marks its first anniversary with membership surpassing 100,000. The…
To build responsible advanced AI, Anthropic is engaging in dialogue with diverse groups worldwide. The first round of…
LlamaIndex engineers released a financial due diligence agent demo app that uses LiteParse to parse SEC filings and…
METR conducted a pilot assessment from February 16 to March 16, 2026, with Anthropic, Google, Meta, and OpenAI…
What are the best practices for running Claude Code at scale? A new blog post shares what we learned from running it…
InsForge is a one-stop open-source backend platform designed for AI coding agents. Through two interfaces, MCP Server…
PaddleOCR released version 3.5, formally establishing Transformers as one of the optional inference backends for…
Today Krea 2 officially opens to everyone. To celebrate the launch, we are offering all subscribers a full week of…
Hugging Face released the Open Agent Leaderboard, which compares complete agent systems rather than just underlying…
Grok can now process video [quote @XFreeze]: you can upload an entire video to Grok and have it analyze, summarize,…
Grok Imagine is now available to all X Premium+ subscribers. This is our image generation model, developed by xAI.
Tencent Cloud has opened public beta of Ardot, its self-developed AI design agent platform. Core features include:…
The Cursor platform released Composer 2.5, with substantially improved intelligence and behavior. The model is better…
xAI officially launched the Grok "Skills" feature on May 18, 2026, aimed at providing persistent specialized knowledge…
Google launched Gemini Omni, a native multimodal AI model that can integrate multiple inputs such as video, images,…
Google Antigravity 2.0 is a new standalone desktop app supporting macOS, Linux, and Windows, with no IDE binding,…
Manus released Scheduled Tasks 2.0, extending scheduling to more contexts including tasks, projects and web apps.…
Anthropic announced the acquisition of Stainless, a developer of SDK and MCP server tools. Since its founding in 2022,…
This article explores the correctness challenges in distributed training and the trade-offs of the DTensor approach.…
Cognition launched Auto-Triage, in which Devin automatically monitors alerts from sources including Slack, Linear,…
Prime Intellect open-sourced general-agent environment v1, which generates a self-evolving task corpus through a…
Google launched the Gemini for Science program, releasing a series of scientific tools and experimental applications…
Page 9 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