StepFun releases Step 3.7 Flash, an efficient Flash model for production-grade agents
StepFun released and open-sourced Step 3.7 Flash, using a sparse MoE architecture (196B+1.8B total parameters, 11B…
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.
StepFun released and open-sourced Step 3.7 Flash, using a sparse MoE architecture (196B+1.8B total parameters, 11B…
This article is the first in a PyTorch profiling series, starting from the simplest matrix multiplication plus bias…
"Skill distillation" is a knowledge transfer method in which frontier large models (such as Opus 4.7, GPT-5.1 and…
Claude Platform on AWS now supports Managed Agents webhooks, multi-agent orchestration, and self-hosted sandboxes. New…
llm-anthropic released version 0.25.1. Key updates include: added the Claude Opus 4.8 (`claude-opus-4.8`) model; added…
DeepSeek has secured its first funding round led by the National Integrated Circuit Industry Investment Fund (the…
Don't rely on benchmarks alone; look at the full picture! Try our new comparison page, which also lets you visualize…
Fine-tuning tests show that large language models carry a bias toward confidently presenting false statements as true,…
Free agentic coding with MiniMax M2.7 on @OpenHandsDev? Yes, please! Available for a limited time ⚡👀
Anthropic has completed a $65 billion Series H funding round at a post-money valuation of $965 billion. It may be the…
hexoai has open-sourced the SIA (Self-Improving AI) framework. The framework shows that AI agents can not only…
Claude Code updated to v2.1.154, officially introducing the Opus 4.8 model. New features include Dynamic Workflows,…
🍌 Nano Banana Pro [gemini-3-pro-image] and Nano Banana 2 [gemini-3.1-flash-image] are now officially released and…
Anthropic released its latest Opus 4.8 large language model and introduced a new tool called Dynamic Workflows for the…
The LMSYS team (Intel and SGLang) enabled a heterogeneous encode-prefill-disaggregated (EPD) architecture for…
SenseNova-U1-8B-MoT-Infographic is an upgraded 8B-parameter infographic generation model. Its core improvements…
Google's Tunix hackathon on Kaggle challenged developers to turn small base models into general reasoning engines…
AI startup Sesame, founded by the Oculus founder, has released its iOS app, bringing conversational AI agents to the…
MiMo-V2.5 is now live on OpenCode — free for a limited time. 🎉 [quote @opencode]: OpenCode x MiMo V2.5 - free for a…
SenseTime introduced its upgraded infographic generation model SenseNova-U1-8B-MoT-Infographic. The model has 8B…
Perplexity Computer is now available in Microsoft Excel, Word, PowerPoint and Outlook. Users can use Computer directly…
Note: You can use Flex and Priority tiers for supported models (OpenAI, Google Vertex, etc.). For pricing information,…
OpenRouter announced $113 million in Series B funding led by CapitalG, with participation from NVentures, ServiceNow…
Mistral AI released the public preview of Search Toolkit on May 28, an open-source composable framework for building…
ElevenLabs has released the Dubbing v2 AI dubbing model, which generates speech directly conditioned on the original…
Mistral has upgraded Le Chat into a unified agent, Vibe, with a single license covering both work and coding, launched…
Exclusive: DeepSeek plans to immediately apply for an IPO on the STAR Market (A-shares) after completing its current…
This tutorial builds a complete pgvector experimental environment in Google Colab, showing how PostgreSQL can serve as…
Linear launched Diffs, a code review product available to all Linear plans starting today, with setup steps in the…
PilotDeck, an agent operating system jointly developed and open-sourced by Tsinghua University's THUNLP lab,…
NVIDIA's research team has open-sourced Polar, an agent reinforcement learning framework. It requires no rewriting of…
Huawei's He Tingbo proposed a new semiconductor evolution path, the "Tao (τ) law", replacing "geometric scaling" with…
Claude Code has released v2.1.153. The main additions are a `skipLfs` option to skip Git LFS downloads and improved…
Jonathan Jaffe, CISO at Lemonade, discusses the new security challenges of the AI agent era. He notes that AI is…
xAI's latest coding model, Grok Build 0.1, has entered public beta via the xAI API. The model is trained for agentic…
vLLM introduced a native weight sync API that standardizes weight transfer from RL trainers to inference engines into…
Easily transform your videos into new visual styles with Gemini Omni. Just upload a video or photo and ask Gemini to…
Bring your own MCP server: your team can keep MCP servers inside your internal network while ChatGPT, Codex, and the…
Krea 2 is now on Comfy! KREA's first foundation image model — trained from scratch — features adjustable creativity,…
Conversation mode has improvements for text and voice input. When a voice session begins, it can access the user's…
Page 2 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