Artificial Intelligence — Page 2
216 articles in this category
OpenAI and Broadcom's "Jalapeño" Chip: Custom Silicon for Inference
OpenAI's first custom chip, Jalapeño, is an inference-focused ASIC co-built with Broadcom, promising better performance-per-watt for large language models.
Xiaomi's HarnessX: Agents That Rewrite Their Own Scaffolding
Xiaomi's HarnessX lets AI agents rewrite their own scaffolding mid-task, delivering a +14.5% average gain, with smaller open models benefiting the most.
Google DeepMind's Gemini for Science Puts AI to Work on Real Discovery
On June 24, 2026, Google DeepMind launched Gemini for Science — AI tools for hypothesis generation, computational discovery, and literature insight, built with top research labs.
ByteDance's Seed 2.1 Models Bring Frontier Coding to a Lower Price Point
ByteDance unveiled Seed 2.1 Pro and Turbo on June 24, 2026 — strong coding and agent models with million-token context and a dramatically lower cost of ownership.
Mistral OCR 4 Turns Documents Into Structured Data You Can Self-Host
Mistral's OCR 4, launched June 23, 2026, reads 170 languages, returns structured layout and confidence scores, tops OCR benchmarks, and runs in a single self-hosted container.
Figma's Config 2026 Adds Motion, an AI-Native Animation Tool on the Canvas
At Config 2026 on June 24, Figma launched Motion — canvas-native, AI-assisted animation — plus Code Layers that turn any design layer into editable code.
Anthropic's New Economic Index Finds Heavy AI Users Feel More Productive and Optimistic
Anthropic's June 26 Economic Index report links AI use to higher self-reported productivity — 86% cite speed gains and most say their skills are growing more valuable.
Gemini's June Drop Adds Live Camera Art and User-Controlled Thinking
Google's June 2026 Gemini Drop brings real-time image generation in Gemini Live and a full global rollout of user-controlled Thinking Levels for everyone.
Qwen-AgentWorld Is an Open Model That Simulates Worlds for AI Agents
Alibaba's Qwen team open-sourced AgentWorld on June 24, 2026 — a language world model that simulates digital environments so AI agents can practice and improve.
AI Finds 1,000+ Antibiotic Candidates Hidden Inside Prion Proteins
Penn's APEX AI scanned prion proteins and surfaced over 1,000 antibiotic candidates on June 22, 2026 — two cleared infections in mice. A genuine AI-for-science win.
OpenAI Launches GPT-5.6 Sol, Its Most Capable AI Model Yet
OpenAI's GPT-5.6 Sol launches as its most capable model yet, leading a new tiered Sol-Terra-Luna lineup with a 'max' reasoning effort and an agentic 'ultra' mode.
GPT-5 Pro Helps an Immunology Lab Crack a Years-Long T-Cell Mystery
On June 24, 2026, OpenAI shared how GPT-5 Pro helped immunologist Derya Unutmaz's lab explain a years-long T-cell puzzle — a real win for AI-accelerated science.
Gemini 3.5 Flash Adds Native Computer Use for Affordable AI Agents
Google DeepMind built computer use into Gemini 3.5 Flash on June 24, 2026, letting developers ship capable, affordable AI agents that see, reason, and act.
PP-OCRv6 Brings Tiny, 50-Language Open-Source OCR to Hugging Face
PaddlePaddle released PP-OCRv6 on June 22, 2026 — a family of tiny open-source OCR models covering 50 languages, now available on the Hugging Face Hub.
OpenAI and Broadcom Unveil Jalapeño, a Custom AI Inference Chip
OpenAI and Broadcom unveiled Jalapeño on June 24, 2026 — OpenAI's first custom AI inference chip, promising better performance-per-watt and major cost savings for everyday AI.
Anthropic's Claude Tag Turns Claude Into a Slack Teammate
Anthropic launched Claude Tag on June 23, 2026, letting teams @-mention Claude in Slack to delegate tasks it works through asynchronously — a shared, 'multiplayer' AI collaborator.
Mistral OCR 4 Brings Private, On-Premise Document AI to Enterprises
Mistral released OCR 4 on June 23, 2026 — a top-scoring document-extraction model that deploys in a single container, keeping sensitive documents inside an organization's own infrastructure.
Google Gemini 2.5 Pro Deep Think Brings Parallel Reasoning to Everyone
Google launched Gemini 2.5 Pro with Deep Think on June 22, 2026 — a parallel-reasoning mode for hard math and coding, with a 2M-token context, now live on the API, AI Studio, and Vertex AI.
Google DeepMind's AI Control Roadmap Charts a Safer Path for AI Agents
Google DeepMind published its AI Control Roadmap on June 18, 2026 — a defense-in-depth framework for safely deploying AI agents that keeps systems secure even when alignment is imperfect.
GLM-5.2 Open Weights Arrive as a Top Coding Model at a Fraction of the Cost
Z.ai released GLM-5.2 open weights under an MIT license on June 16, 2026 — an open-weight coding model that rivals the best closed systems on long-horizon benchmarks at roughly one-sixth the cost.
Anthropic Opens a Seoul Office and Deepens Its Korean AI Partnerships
Anthropic opened its Seoul office on June 17, 2026 — its third in Asia-Pacific — alongside a wave of enterprise, academic, and non-profit partnerships bringing Claude across Korea's tech ecosystem.
Grok Build's New Dashboard Lets Developers Run Coding Agents in Parallel
xAI's June 17 Agent Dashboard for Grok Build puts every coding session on one screen, surfacing the ones that need input so developers can orchestrate multiple AI agents at once.
OpenAI Partner Network Invests $150M to Scale Enterprise AI Adoption
OpenAI launched its $150M Partner Network on June 14, aiming to certify up to 300,000 consultants and help enterprises deploy AI through tiered systems-integrator partnerships.
Moonshot's Kimi K2.7 Code Arrives as an Efficient Open-Weight Coding Model
Moonshot AI's open-weight Kimi K2.7 Code launched June 12 with a 1T-parameter MoE design, a 256K context window, and roughly 30% lower reasoning-token use.