agent · 317 topics

vulnerability in agents from google and others exposes structural flaw in mcp

vals ai deploys 90 claude agents to hunt room temperature magnetic semiconductors

contrastive language models

thinkingbox

cloudflares clef beats rivals with 38ms typed decisions

cerebras qwen3 5 agent benchmark

cognition slashes devin review costs by 70 with smarter model routing

h company holo4 computer use

taichuai zdtaichu5 0 9b multimodal foundation model

fastino gliner2 5 decide

stanford clm turns agent decisions into vector search 9x faster

xiaomi mimo v2.6 distill qwen 9b

xai ships grok 4 7

alibaba qwen3 8 omni flash

googles dream rsi cuts ai discovery agent calls by 162x using replay

shared selective persistent memory for agentic llm systems

your agent aced the task will it do it again

artificial analysis capability indices v1 1

tech watch ai infrastructure

wikiskill compiling agent experience

nex n2.5 pro

deepseek ai deepseek v4 1 flash

openai gpt 6 astra release

meta muse spark 1 3

7 ai models autonomous businesses

spark x25 4b release

cohere labs ate dataset reveals only 2 6 of ai agent tools actually work

openais rogue agents were caught communicating via public wikis

openai gpt 6 astra computer use

mbzuai k2 horizon 375b

gemini 3 8 flash and 3 8 flash cyber

metas muse spark 1 3 cuts tool calls

introducing agentic video understanding with gemini

huggingface breach openai metr report

apodex 1 1 beats deepseek on real world agentic work with a 35b open model

perplexity adds glm 5 3

sapiens ai agnes 2 5 pro beta

anthropic automated alignment researcher

epochs ebr bench reveals humans learn while frontier ai stays stuck

agent seer synthesizing scenarios

perplexitys brain beats vector search with a self rewriting memory wiki

1200 openai agents broke out of sandboxes and hacked hugging face

granite 4 2 llms

30b open models trend

nvidia harness arc agi 3 breakthrough

anthropic agent stack batch actions

ornith 1 5 35b a3b

glm 5 3

how much memory does your agent actually need

import ai 469 dig bench faraday

nvidia nemotron rl agentic terminal pivot v1

prime intellect fable 5

muse glimmer 30b gguf

introducing gemini 3 7 flash

deepseek v4 pro release

perplexity search as code benchmarks

spacexai grok 46

upstages solar pro 4 jumps 28 points to beat human agents

altk evolve fewer tokens

muse glimmer

liquidai lfm2 5 2 6b

meta muse code and spark 1 2

prime agent beats human experts on arc agi 3

microsoft orchard framework

xyz ailab xyz aquila mini

microsoft fara1.5 27b

echoverse deep evolving environments for computer use agents

handbook md long policy documents agents

claude opus 5 vending bench

kwaipilot kat coder v2 5 dev

poolside laguna s 2.1

upstage solar open2 250b

nanbeige nanbeige42 3b

glint research fable 5 traces

claude fable 5 traces

context rot is a harness problem and rlms show why

openai sandbox escape

gpt 5 6 claude fable 5 muse spark 1 1

what building shippy taught us about building agents

proactive agent research environment

glm 5 2 vat benchmark

data for agents

nvidia bionemo agent toolkit

gemini 3 5 flash computer use

google gemini 3 5 flash computer use

qwen agentworldbench

plamo 3 0 prime

from the hugging face hub to robot hardware with strands agents and lerobot

is it agentic enough

holo31 fast local computer use

anthropic claude opus 4 8 release

itbench aa frontier models score below 50 percent

2026 05 29 papers 2605.27375

microsoft fara 7b cua

hermes agent reasoning traces

2026 04 27 papers molclaw 2026 04 02

2026 04 27 papers 2604 22085

qwen qwen3 6 35b a3b trending

kimi k2 6 multimodal agentic model

anthropic claude sdks browser computer toolset

docker agent

reflection ai beam

fine tune a search agent with multi turn rl on amazon sagemaker ai

pi 1 0

openai codex cli rebuilds the terminal to run parallel coding agents

meta muse azure openai model

github copilot canvas custom ui

anthropic claude art

ax google open agentic orchestrator

cognition adds xai grok 4 7 to devin

openjev agent decision acceleration

sub second artifact deployments are now supported in vercel cli

devin macos vm

ai sdk harness layer native subscription authentication

anthropic mcp david soria parra keynote

agent team multibot

reef continual learning infrastructure

nex agi nex n2___5 mini

agent evaluation metric for multi turn conversations

openclaw 9agents management os

chat agent deprovisioning design

claude in chrome ga

agent agency huggingface incident

openai gpt 6 astra openrouter

grok bot

run agent driven amazon sagemaker hyperpod operations with instantstart

grep beats lsp

cursor self hosted agents

anthropic commerce agents blueprint

claude code mac background computer use

github copilot efficiency

nous research hermes agent pantheon

perplexity ships pii tracer

agent memory as a file format

optimistic parallel orchestration for ai agents

github copilot slack teams

ai review planted bugs calibration

anthropic model hardware standard

run claude managed agents with chat sdk

cad 1000 hours

coding agent harness source code analysis

model context protocol roadmap 2026

cursor origin

design ai systems with provisional agents

microsoft agent framework for go

mole deep research agent

nvidias nemo switchyard cuts agent ai costs by 74

nous research hermes agent loop command

deepseek harness developer preview

artificial analysis optima

llama cpp

nous research hermes agent profiles

oneadvanced deployed over 50 ai agents on uk sovereign aws

meta models muse glimmer 30b

first orion amazon nova act

openchamber

mcp agentcore gateway checklist

anthropic claude code cross session messaging

cloudflare os launch

devtools must be open source

cloudflare computer

vercel v0 api v2

rpr ai governance runtime

datasette apps 0 2a0

deepseek v4 flash public beta

agentic catalog experience amazon quick

amazon bedrock agentcore observability performance optimization

bolt new security agent

moonshotai kimi k3

anthropic rebuilds mcp as stateless

modal cto clarifies sandbox incident

claude cookbook

laguna s 2 1

drawing the mona lisa with gpt 5 6 claude gemini and grok

agent swarms and the new model economics

setting up your spare mac for claude code

lm studio bionic

how smartsheet built a remote mcp server on aws

the llm critics are right i use llms anyway

i tricked claude into leaking your deepest darkest secrets

towards a harness that can do anything

rl agent trains ai

what will be left for us to work

the agentic loop

migrating a production ai agent to gpt 5 6

claude code vs opencode token overhead

claude code sandbox

who manages the agents

automating cross repo documentation with github agentic workflows

rowboat ai desktop coworker

ai pr automation codens

qwen agentworld 35b a3b

claude fable 5 autonomous agent

llm coding ux experimentation

senior swe bench

zcode claude code from the makers of glm

zcode harness for glm 5 2

ornith 1 0

hatcha

claude code extended thinking summary only

a robot is sprinting towards you do you want it running on claude or grok

open code review

anthropic vulnerability discovery framework

github copilot app

dynamic workflows in claude code

kanbots parallel agents kanban

qwen3 7 max agent frontier

forge guardrails

claude code large codebases

anthropic finance agents

open design

openai symphony orchestration spec

gpt 5 5 prompting guide

gpt 6 astra starcraft cheat

openai dots agent

openai dots agent

metas ai tamagotchi bet

microsoft copilot os

perplexity computer adds minimax and seedance

anthropic claude chat cowork integration

gennai gov ai agent

xai grok bot enterprise

plaud one ai earbuds

google wires gemini live into spark

chatgpt computer history

google maps agentic features

viktor team agent

esol ai mobility sandbox

chatgpt work agent password protected sites

openais new voice mode makes it to the chatgpt desktop app

buzz group chat platform

anthropic claude cowork record skill

google expands gemini spark

openai codex micro hardware

siri ai ios 27 public beta

claude cowork expands to mobile and web

openclaw dating automation

google gemini spark 24 7 agent

google gemini 3 5 flash ai search update

google io 2026 gemini omni antigravity

how chatgpt adoption broadened in early 2026

livemacebench process aware evaluation llm agents

memadapter counterfactual adaptation

tracedance deployment traces agent benchmarks

compoworld compositional environment scaling

just in time memory llm agents

llm agents can easily tamper with their own traces

pistis technical report

training capable and safe computer use agents

scienceide scientific codebase agent environments

modularrsi modular recursive harness self improvement

atria dawn the dawn of agentic superintelligence

terminal agent reinforcement learning for long horizon tasks

omni interaction agent technical report

cost aware evaluation of long term memory in tool using llm agents

autofyn non parametric expert iteration

terminal universe

harbor adapters and harbor index

speculative macro commit

jit agent harness evolution

frontierchallenge evaluating scientific workflow completion

icu mortality llm agentic pipeline

wikiskill compiling agent experience into persistent knowledge

spade self play adaptive synthetic executable environments

fm bench

ogx open source vendor neutral generative ai application server

proxy validated llm ux micro simulations

rubricforge agent evaluation

spatial memory agent

spark to paper

llm agents factory retrieval of domain specific llm agents

cyberforge repository level vulnerability injection

economic world models systems blueprint

video deepresearch

recursive synthetic terminal tasks

toolartist

worldclaw agentic 3d generation

kernelbrain coarse to fine budget aware search for agentic gpu kernel optimization

memarena ego centric benchmark on device agentic memory

videococo code as cot

magicselector joint optimization for agent tool selection

vld rag agentic multimodal rag long documents

craft learn the schema execute the plan

deeplens diagnosis agent

sf ams structured memory llm agent

arex recursively self improving agent deep research

nvidia labs oo agents

solar open 2 technical report

rail guard closing the evaluation to remediation gap

mcpevol bench benchmarking llm agent performance across dynamic evolutions of mcp servers

retroagent harnessing llms to search over structured memory

eta given delta

deepsearch world

agentlens

memora harmonic memory

qwen agentworld

rift bench dynamic red teaming

the harness is the company

openai agent used dns to reach external chatbot

openai bots meddled with us government sites

openai pauses training

openai agent australian government breach

openai agent hacked australian government website

aiuc series a

baseten github pat takeover

openai agent research acceleration safety

openai agents sandbox escape wiki

open executive

figure ai brett adcock bets gigawatt nvidia deal on hark handoff

metr raises 71m to independently stress test ai

manus meta independence

openai astra critical cyber status

openai astra model pause

openai active users 1b 2m enterprise

openai rogue ai agent hacked more than hugging face

prentis ai lab funding

ai agent startup fundraise