Keeping the Cursor app stable
The Cursor team focused on out-of-memory crashes, given users rely on the app all day and crashes have severe impact.…
AI news monthly archive: April 2026 features 272 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 Cursor team focused on out-of-memory crashes, given users rely on the app all day and crashes have severe impact.…
QIMMA is an Arabic large language model evaluation platform that introduces the first quality verification process. It…
The article analyzes the complex factors behind a single evaluation number for open-source and closed-source AI…
Kimi released version K2.6, focusing on improving open-source coding capabilities. The new version optimizes code…
METR classified 77 records in the NanoGPT speedrun to assess AI agents' contribution to AI R&D: before March 2026,…
DR-Venus is a 4-billion-parameter deep research agent trained with only 10,000 open data samples, based on the…
The UK AI Security Institute deployed the open-source coding agent OpenClaw in a sandbox and prompted it to explore…
Anthropic Labs launches Claude Design, a visual design tool powered by Claude Opus 4.7, open as a research preview to…
LLaDA2.0-Uni is a unified multimodal model with world understanding and generation capabilities. By integrating…
AutoClaw officially launched its self-evolution mechanism and Skill Store. After each conversation, it scans for user…
xAI released two standalone APIs, Grok speech-to-text (STT) and text-to-speech (TTS), using the same technology stack…
The CRUX project, comprising 17 researchers, tests AI capabilities in real-world scenarios through 'open-world…
Chrome today upgrades its AI Mode feature, using artificial intelligence to deeply change how users interact with web…
Nano Banana 2 now supports generating images by combining personal context with Google Photos data, analyzing users'…
A2UI v0.9 released a new framework-agnostic standard designed to help AI agents generate customized UI components in…
Claude Opus 4.7 is fully rolled out, achieving a major leap in advanced software engineering tasks. Its solve rate on…
Qwen open-sourced Qwen3.6-35B-A3B, a sparse MoE model with 35 billion total parameters and only 3 billion activated…
The Codex app has launched a major update, officially supporting both macOS and Windows. The new version integrates…
MiniMax officially launched MaxHermes, the world's first cloud sandbox AI assistant built on Hermes Agent. It…
OpenAI has released GPT-Rosalind, a specialized reasoning model for life sciences. The model is designed for drug…
Ecom-RLVE is an adaptive verifiable environment designed for e-commerce conversational agents. The environment…
As AI code agents matured in 2026, open-source projects such as transformers faced a surge in pull requests of…
Sentence Transformers released capabilities for training and fine-tuning multimodal embedding and reranking models.…
Anthropic released Claude Opus 4.7, positioned as its most capable general-purpose model, excelling at complex…
TensorZero's measurements found that token counts for the same input can differ by up to 2.65 times across vendors,…
Jensen Huang said Nvidia already has a supply chain system capable of supporting a trillion-dollar-scale business over…
Google has officially released Gemini 3.1 Flash TTS as the next generation of expressive AI voice technology, and the…
A blog post from IBM Research published on Hugging Face takes a detailed look at the core mechanisms of its agent…
An eight-month study of developers using Cursor at 500 companies found that after the release of advanced models such…
OpenAI has launched a major update to the Agents SDK, introducing a native sandbox execution environment and a…
HCompany published a blog post on Hugging Face officially launching HoloTab, an AI browser companion. Positioned as an…
Gemini CLI introduced subagents, specialized expert agents that handle complex or high-throughput tasks in independent…
The OpenRouter platform has launched a video generation service. Users can directly call top video generation models…
Cursor added a Canvas feature that turns information into visual, interactive interfaces, replacing long, hard-to-read…
Modal has released an engineering tutorial demonstrating how to build a custom coding agent framework from scratch…
V8.1 Alpha has been officially released, succeeding the V8.0 model tested for the past month. This version returns to…
Claude Code has released a research preview of routines, supporting automated workflows that include prompts, code…
Chrome has added a Skills feature that lets users discover, save, and recombine AI workflows for instant one-click…
Mistral released Connectors in Studio as a public preview. Built-in connectors and custom MCP can be used via API/SDK…
ByteDance Seed released the native multimodal audio-video generation model Seedance 2.0, which officially launched in…
Page 5 of 7
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