AI Agents
AI agents are LLM-powered systems that don't just answer — they plan steps, call tools and drive a task to completion: writing code, searching the web, booking, analyzing. 2026 became the year of agents, from coding agents to autonomous researchers. This page gathers all our coverage of agentic AI: launches, protocols (MCP, A2A), reliability and real-world use.
Moonshot Releases WorldVQA: 3500 Questions About the World That Trip Up Models
The Kimi team published WorldVQA, a benchmark of 3500 image-question pairs. On rare objects, top multimodal models often fall short of 50% accuracy.
Kimi launches Agent Swarm: up to 100 subagents and 1,500 tool calls
Moonshot built a swarm mode into Kimi K2.5: the model hires up to 100 subagents on its own, makes more than 1500 tool calls and, the company says, works 4.5x faster.
Moonshot open-sources Kimi K2.6: 300 sub-agents and 13 hours of autonomous coding
Moonshot open-sourced Kimi K2.6: the model spent 13 hours tuning an exchange engine, sped up Zig inference from 15 to 193 tokens/s, and scales its swarm to 300…
OpenAI shows how an engineer's dot triaged a night-time incident and launched Codex
Google DeepMind Launches AI Agent to Hunt for Antimicrobial Peptides Among Millions of Proteins
OfficeTown Simulates 23 Organizations and Trains Agents for Office Work on 406 Tasks
Reflection releases Beam: 501B parameters and 30 points on the Intelligence Index
ERB: AI Agents Build Their Own Model Tests and Match the Reference Benchmarks in 75% of Pairs
The EvalResearchBench benchmark tested 9 research agents. Their evaluators rank models the same way as the target benchmarks in 75% of pairs, against a ceiling of…
RoboHarness: A Coding Agent on Qwen3.8-Flash-Next Beats VLA Models in BEHAVIOR
Researchers at Zhejiang University have unveiled RoboHarness, a visual-geometric control panel that lets an LLM agent outperform VLA models on BEHAVIOR Challenge…
Same 74 Tasks, Different People: Browser AI Agents Lose 36% of Their Success Rate
Codex Teaches Team Sites to Update on a Schedule and Adapt to Each Visitor
OpenAI Releases GPT-Live-1.1: New Voices in 6 Languages and Fewer "checking..." Pauses
EinsiaAI Releases EmbodiedEvo: Self-Evolving Agent Tops Robodojo
Vera Rubin NVL72 beats GB300 by up to 5.7x in tokens per megawatt in AA-AgentPerf
Artificial Analysis published the first Vera Rubin NVL72 results in its agentic coding benchmark: up to 5.7x throughput per megawatt over GB300 and a record 906…
Cohere pitches utilities on running North AI agents inside their own perimeter
Cohere's new guide for utilities argues that AI in critical infrastructure hinges less on model capability than on where it runs, what data it touches, and what…