I still remember the excitement in 2023 when Stanford Smallville was launched. It was the largest mu...
I still remember the excitement when Stanford Smallville was launched in 2023. It was the largest multi-agent…
AI news monthly archive: January 2026 features 51 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.
I still remember the excitement when Stanford Smallville was launched in 2023. It was the largest multi-agent…
Flood is an efficient large language model inference framework for offline applications. It uses pipeline parallelism…
Anthropic trained three explicitly sabotaging models based on Claude Sonnet 4.5 Internal and ran blind audit tests…
Structured outputs for the Claude API are now fully available for Claude Sonnet 4.5, Claude Opus 4.5 and Claude Haiku…
OpenAI announced it will retire GPT-4o, GPT-4.1, GPT-4.1 mini and OpenAI o4-mini in ChatGPT on February 13, 2026, in…
xAI launched the Grok Imagine API, offering text/image-to-video and high-precision editing, with support for adding or…
Arcee AI released the open large model Trinity Large, marking a full commitment to building open models in the United…
In the era of agents, management ability will become the core superpower for humans to thrive. In an AI-dominated…
LlamaIndex has released the beta of LlamaAgents Builder, which lets users generate document processing agents based on…
Hugging Face published a blog post discussing core architectural choices and development paths in China's open-source…
Kimi released and open-sourced K2.5, its most intelligent and versatile model to date. Its native multimodal…
The LinkedIn team explored the feasibility of using the GPT-OSS model as the core of agent applications for…
PromptArmor published research demonstrating an unfixed vulnerability in Claude Cowork where hidden prompt injection…
Dario Amodei published a long essay, "The Adolescence of Technology," noting that powerful AI is in its "adolescence"…
Mistral released Mistral Vibe 2.0, bringing custom sub-agents, multi-option clarification, slash command skills, a…
MiniMax released its next-generation voice model MiniMax Speech 2.8, which uses native sound tag technology to…
OpenAI's developer blog has published a tutorial on how to systematically evaluate Codex's Agent skills with evals,…
AI agents are approaching a critical threshold, and their performance leaps exceed what traditional work models can…
Kimi open-sourced the Kimi Vendor Verifier (KVV) alongside the K2.5 model release to verify the accuracy of…
The lead authors of METR's time horizon paper published a clarification of common misreadings of the metric, stating…
IBM Research released AssetOpsBench on Hugging Face, an AI agent benchmark framework for industrial asset operations…
PromptArmor disclosed a data exfiltration vulnerability in OpenAI platform 'responses' and 'conversations' API logs…
Tristan Hume, head of Anthropic's performance optimization team, shared experience in designing AI-resistant technical…
Cognition released Devin Review, an AI code review tool for GitHub PRs, saying code review has become the bottleneck…
Differential Transformer V2 has been released, centered on an improved differential attention mechanism. Compared with…
Overworld introduced Waypoint-1, a real-time interactive video diffusion model that lets users control generated,…
PromptArmor released a secure deployment guide for Claude Cowork, outlining its threat model and offering three…
Liquid AI released LFM2.5-1.2B-Thinking, a reasoning model that runs fully offline on phones with about 900 MB memory…
We have officially open-sourced Baichuan-M3, a new-generation medical large language model, which scored 65.1 on…
OpenAI announced that ChatGPT Go is now available globally, offering expanded access to GPT-5.2 Instant, higher usage…
OpenAI plans to test ads in ChatGPT Free and Go in the US to expand affordable global access to AI. The company says…
OpenAI is advancing the democratization of artificial intelligence through open source and open science. Its core goal…
The research team released an experimental dataset called CoVal containing crowdsourced evaluation rubrics. The…
LlamaIndex argues that agents increasingly read unstructured context through files and file systems, summarizing three…
OpenAI announced a partnership with Cerebras to add 750MW of high-speed AI compute to reduce inference latency and…
Meituan LongCat launched Heavy Thinking Mode based on the 560-billion-parameter MoE large model…
PromptArmor disclosed an indirect prompt injection vulnerability in Superhuman AI where hidden instructions in a…
The Manus blog published an evaluation of the top 10 prompt-to-video AI generators for 2026, testing Runway (Gen 4.5),…
Anthropic proposes a "confessional" training method that requires AI to internally generate a safety explanation to…
Agent-of-empires (AoE) is an AI coding agent session manager supporting Linux and macOS, compatible with 9 mainstream…
Page 1 of 2
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