AI Research
New MRAgent Framework Uses 118K Tokens Per Query, Outperforms LangMem
MRAgent framework reduces token consumption and runtime costs for long-horizon reasoning tasks in AI agents.
VentureBeat
Jun 26, 2026·1 min read
Source
VentureBeat
60 articles from this source
AI Research
Autonomous Security Agents Require Complete Data: Is Yours Ready?
Incomplete data hampers autonomous security agents; experts stress verifying endpoint coverage before deployment.
VentureBeat
Jun 26, 2026·1 min read
AI Models
OpenAI unveils GPT-5.6 family with Sol, Terra, and Luna models
OpenAI announces limited preview of GPT-5.6 family, including Sol, Terra, and Luna models for various enterprise needs.
VentureBeat
Jun 26, 2026·1 min read
Productivity
Most companies think they're building a software factory. They're actually just shipping bugs faster.
Industrialized factories changed how the world produced physical goods: more output, lower costs, faster than anything that came before.
VentureBeat
Jun 26, 2026·1 min read
AI Models
Liquid AI releases smallest AI model yet, LFM2.5-230M, for data extraction and local deployment
Liquid AI's LFM2.5-230M model outperforms larger models in data extraction and can run on local devices.
VentureBeat
Jun 25, 2026·1 min read
AI Models
OpenAI's GPT-5.5 Instant Model Updated with Improved Shopping and Constraint Handling
OpenAI updates GPT-5.5 Instant model, used in free ChatGPT version, with improved shopping results and complex constraint handling.
VentureBeat
Jun 25, 2026·1 min read
AI Startups
Mindstone's Rebel Automates Enterprise AI Model Selection
Mindstone's Rebel enables enterprises to automatically select the best AI model for each task and subtask, ensuring reliable and secure AI workflows.
VentureBeat
Jun 24, 2026·1 min read
AI Models
Mistral Launches OCR 4, Pushing Document Extraction Beyond Text
Mistral AI releases OCR 4, a document intelligence model that extracts structured representations of documents, complete with bounding boxes and confidence scores.
VentureBeat
Jun 24, 2026·1 min read
AI Research
Alibaba's Qwen-AgentWorld Improves Agent Performance Across Seven Benchmarks
Alibaba's Qwen team releases Qwen-AgentWorld, two models trained to predict environment returns, boosting agent performance in seven domains.
VentureBeat
Jun 24, 2026·1 min read
AI Research
Xiaomi's HarnessX Boosts AI Performance by Evolving Software Scaffolding
Xiaomi's HarnessX framework autonomously improves AI software scaffolding, yielding significant performance gains across domains.
VentureBeat
Jun 24, 2026·1 min read
AI Research
Stanford Researchers' Agentic AI Scientists Set to Reshape Drug Discovery
Stanford researchers develop agentic AI 'scientists' to streamline drug discovery, reducing inefficiency and high failure rates.
VentureBeat
Jun 24, 2026·1 min read
AI Tools
How Shopify built an AI stack that doesn't care which models survive
Shopify built an LLM proxy that gives every engineer access to multiple AI providers — with automatic failover when any one of them goes down, changes, or disappears.
VentureBeat
Jun 24, 2026·1 min read
AI Research
Amazon to present framework for engineering trustworthy AI agents at VB Transform 2026
Amazon's AGI autonomy lab develops framework for trustworthy AI agents, focusing on consistency, robustness, and safety.
VentureBeat
Jun 24, 2026·1 min read
Big Tech
Intuit to showcase AI infrastructure rebuild at VB Transform 2026
Intuit overhauled its AI infrastructure to support complex tasks, shifting from a multi-agent setup to a granular, skill-and-tool-based architecture.
VentureBeat
Jun 24, 2026·1 min read
AI Models
OpenAI and Broadcom unveil custom AI inference chip Jalapeño
OpenAI and Broadcom partner on Jalapeño, a custom AI accelerator chip for large language model inference.
VentureBeat
Jun 24, 2026·1 min read
Image AI
Krea Releases Open-Weights AI Image Models with 2-Second Generation Speed
Krea releases Krea 2 Raw and Turbo, open-weights AI image models with fast generation speeds, under a custom license.
VentureBeat
Jun 23, 2026·1 min read
AI Tools
Anthropic Launches Claude Tag, an AI Teammate for Slack
Anthropic launches Claude Tag, an AI agent that embeds directly inside Slack as a persistent teammate that anyone can delegate work to.
VentureBeat
Jun 23, 2026·1 min read
AI Research
AI Data Delivery: The Key to Scaling Reliable Production Workloads
Enterprises struggle to scale AI workloads due to fragile data delivery infrastructure.
VentureBeat
Jun 23, 2026·1 min read
Video AI
Alibaba's AI video model rises to No. 2 in global rankings, as OpenAI's Sora and ByteDance's Seedance fall away
Alibaba Cloud on Sunday released HappyHorse 1.1 , a major upgrade to its AI video generation model that the company says delivers production-ready video synthesis across core content creation scenarios.
VentureBeat
Jun 22, 2026·1 min read
AI Startups
Sakana Launches Fugu, a Multi-Agent Orchestration System for Frontier AI Performance
Sakana AI launches Fugu, a multi-agent orchestration system delivering frontier-level AI performance through a single, OpenAI-compatible API.
VentureBeat
Jun 22, 2026·1 min read
AI Research
Agentic Enterprises Must Become Learning Systems to Stay Ahead
Organizations need to capture and reuse knowledge gained from daily operations to improve AI-driven decisions.
VentureBeat
Jun 22, 2026·1 min read
AI Research
Researchers Introduce Self-Harness, a Framework for AI Agents to Rewrite Their Own Rules
Self-Harness lets AI agents systematically improve their operating rules, boosting performance up to 60% without relying on human engineers or stronger external models.
VentureBeat
Jun 22, 2026·1 min read
AI Research
AI hits memory wall, now needs new context tier
As AI inference workloads evolve, GPU availability is no longer the primary bottleneck; instead, context management has become a major challenge.
VentureBeat
Jun 22, 2026·1 min read
AI Tools
7,000 Langflow servers are under attack. LangGraph and LangChain have the same holes
Your AI agent did exactly what it was designed to do.
VentureBeat
Jun 19, 2026·1 min read
AI Research
AI Agents Stumble in Production: Can Hypernetworks Offer a Solution?
Enterprise AI agents often stall in production, requiring human oversight, but hypernetworks may offer a solution by generating task-specific models on demand.
VentureBeat
Jun 19, 2026·1 min read
AI Tools
Anthropic's Claude Code Artifacts update brings live, shared dashboards and interactive workspaces to enterprises
Anthropic announced a potentially game-changing new feature for users of Claude Code on the Claude Team and Enterprise subscription plans: Artifacts .
VentureBeat
Jun 18, 2026·1 min read
AI Models
New AI optimization framework Arbor outperforms Claude Code and Codex by 2.5x
Arbor framework automates AI-driven research and optimization, outperforming Claude Code and Codex by 2.5x on the same compute budget.
VentureBeat
Jun 18, 2026·1 min read
AI Tools
Copilot searched your mailbox. LiteLLM handed out admin keys. Run this 5-check audit before your stack is next
Two AI tools broke in the same way in the same two weeks, and four research teams proved it.
VentureBeat
Jun 18, 2026·1 min read
AI Tools
Adobe Embeds AI Workflows Across Creative Cloud, Shifting from Media Generation to Production Orchestration
Adobe expands 'creative agent' across Creative Cloud suite and upgrades Firefly AI studio to automate complex production workflows.
VentureBeat
Jun 18, 2026·1 min read
AI Research
AWS enters context layer market with graph that learns from agents
AWS launches context intelligence stack with knowledge graph service that improves over time through agent usage.
VentureBeat
Jun 17, 2026·1 min read
AI Tools
Anthropic ships major Claude Design overhaul with design system imports, code round-trips, and a fix for its token-burning problem
When Anthropic quietly released Claude Design in April as a " research preview ," it generated the kind of instant traction most product teams dream about: more than one million users in its first week.
VentureBeat
Jun 17, 2026·1 min read
AI Research
Weibo's VibeThinker-3B sparks debate over AI benchmarks and scaling laws
A 3 billion parameter language model from Sina Weibo achieves benchmark scores comparable to much larger models, sparking debate over AI benchmarks and scaling laws.
VentureBeat
Jun 17, 2026·1 min read
AI Models
Z.ai's GLM-5.2 Open-Weights Model Beats GPT-5.5 on Coding Benchmarks at Fraction of Cost
Chinese AI startup Z.ai releases GLM-5.2, a 753-billion parameter open-weights LLM that outperforms GPT-5.5 on multiple long-horizon coding benchmarks at 1/6th the cost.
VentureBeat
Jun 16, 2026·1 min read
AI Startups
Databricks claims to have solved decades-old data pipeline problem hindering AI agents
Databricks announces two products to unify operational and analytical databases, eliminating latency and performance degradation.
VentureBeat
Jun 16, 2026·1 min read
AI Research
Stanford's DeLM Cuts Multi-Agent Task Costs by 50% Without Central Orchestrator
Stanford's DeLM framework enables agents to coordinate directly, reducing multi-agent task costs by 50% without a central controller.
VentureBeat
Jun 16, 2026·1 min read
Big Tech
Satya Nadella warns AI could hollow out industries, echoing globalization damage
Microsoft CEO Satya Nadella warns AI could concentrate value, commoditize industries, and urges businesses to build proprietary learning loops.
VentureBeat
Jun 15, 2026·1 min read
AI Startups
Sakana AI Launches 'Ultra Deep Research' Agent for 100+ Page Reports in 8 Hours
Tokyo-based Sakana AI debuts Sakana Marlin, an autonomous B2B research agent generating in-depth strategy reports.
VentureBeat
Jun 15, 2026·1 min read
AI Research
85% of IT teams claim every AI agent is under control. Only 42% actually know who owns them.
Organizational leaders are nearly twice as likely to hide their AI use compared to all other employees, at 42% versus 23%, according to new Ivanti research surveying 3,900 employees across six countries.
VentureBeat
Jun 15, 2026·1 min read
AI Research
AI-powered deception forces defenders to prioritize truth at machine speed
AI has changed the economics of cyber deception, making it essential for defenders to prioritize truth at machine speed.
VentureBeat
Jun 15, 2026·1 min read
AI Models
Anthropic blocks public access to Claude Fable 5 and Mythos 5 models following US government order
US government issues export control directive, citing national security, for Anthropic to suspend access to top-tier AI models for foreign nationals.
VentureBeat
Jun 13, 2026·1 min read
AI Models
Moonshot AI's Kimi K2.7-Code update claims 30% reduction in thinking tokens
Moonshot AI releases Kimi K2.7-Code, an open-source update to its K2 coding model family, with claimed performance gains and reduced thinking tokens.
VentureBeat
Jun 12, 2026·1 min read
AI Research
Google researchers introduce 'faithful uncertainty,' allowing LLMs to offer best guesses instead of hallucinations
Large language models continue to struggle with hallucinations, presenting a major roadblock for real-world enterprise applications.
VentureBeat
Jun 12, 2026·1 min read
AI Tools
NanoClaw and JFrog Partner to Block AI Agents from Downloading Malicious Code
NanoClaw and JFrog launch joint security integration to protect AI agents from malicious code injection.
VentureBeat
Jun 12, 2026·1 min read
AI Tools
PixelRAG outperforms text parsers in accuracy and cuts AI agent token costs
PixelRAG skips text parsing, rendering web pages as screenshots to improve retrieval accuracy and cut AI agent token costs by 10x.
VentureBeat
Jun 12, 2026·1 min read
AI Research
Microsoft's SkillOpt framework optimizes AI agent skills without changing model weights
Microsoft's open-source SkillOpt framework helps AI agents adapt to new domains by optimizing their skills without changing the underlying model weights.
VentureBeat
Jun 11, 2026·1 min read
AI Models
Xiaomi's MiMo Code Outperforms Claude Code on Long-Horizon Coding Tasks
Xiaomi's open-source MiMo Code V0.1.0 beats Anthropic's Claude Code on agentic coding benchmarks, especially on 200+ step tasks.
VentureBeat
Jun 11, 2026·1 min read
AI Research
AI benchmarks overlook real-world performance issues
AI teams focus on compute and storage, but neglect network issues that cause performance drops in production.
VentureBeat
Jun 11, 2026·1 min read
AI Models
Google Releases DiffusionGemma, a Diffusion-Based Language Model
Google's DiffusionGemma generates 256 tokens in parallel, self-correcting as it goes, with speeds up to 4x faster than standard models on GPUs.
VentureBeat
Jun 11, 2026·1 min read
AI Research
Why AI that works in the lab often fails in production — and what actually fixes it
Enterprises struggle to make AI work in the real world, not to experiment with it.
VentureBeat
Jun 11, 2026·1 min read
AI Research
GPT-5.5 beats Claude Fable 5 on Agents' Last Exam benchmark
Researchers launch Agents' Last Exam, a benchmark testing AI's ability to execute long-horizon professional workflows, with GPT-5.5 taking top spot.
VentureBeat
Jun 10, 2026·1 min read
AI Research
Researchers Train Foundation Model from Scratch for $1,500
Sapient's HRM-Text model achieves competitive performance with much larger models at a fraction of the cost and training data.
VentureBeat
Jun 10, 2026·1 min read
AI Research
Anthropic CEO Calls for FAA-Style Regulation of Powerful AI Models
Anthropic CEO Dario Amodei calls for government regulations on powerful AI models, comparing the industry to commercial aviation.
VentureBeat
Jun 10, 2026·1 min read
Productivity
MassMutual's AI strategy focuses on flexibility and measurable outcomes
MassMutual uses 12-month contracts, measures productivity gains, and avoids vendor lock-in to stay agile with AI.
VentureBeat
Jun 10, 2026·1 min read
Big Tech
Apple’s new Siri AI is more than just a smarter assistant — it's a new enterprise app layer
Apple’s new Siri AI, unveiled yesterday at Apple's annual Worldwide Developers Conference (WWDC 2026), may look like a consumer product story on the surface.
VentureBeat
Jun 09, 2026·1 min read
AI Tools
Cohere Open-Sources North Mini Code, a Coding Agent for Agentic Pipelines
Cohere releases North Mini Code, an open-source coding agent that runs on a single H100, targeting agentic software engineering and coding pipelines.
VentureBeat
Jun 09, 2026·1 min read
AI Research
On-device AI agents hit a hard memory limit. Apple's new architecture routes around it.
On-device AI models have stayed small because the entire weight set has to live in DRAM, capping practical parameter counts well below what server-side deployments use.
VentureBeat
Jun 09, 2026·1 min read
AI Models
Anthropic Launches Claude Fable 5 and Claude Mythos 5 AI Models
Anthropic releases Claude Fable 5 and Claude Mythos 5, its most powerful generally available AI models, with enhanced performance and safeguards.
VentureBeat
Jun 09, 2026·1 min read
AI Research
Open-source AI search agent Harness-1 outperforms GPT-5.4 on recalling relevant information
Researchers develop Harness-1, a 20-billion parameter open-source search agent that surpasses GPT-5.4 in recalling relevant information.
VentureBeat
Jun 08, 2026·1 min read
AI Research
Agentic AI solved coding — and exposed every other problem in software engineering
The integration of agentic AI in software engineering has accelerated code generation, but also revealed deeper challenges in defining requirements, integrating complex systems, and maintaining software under real-world conditions.
VentureBeat
Jun 07, 2026·1 min read
AI Tools
When Claude changed, everything changed: Managing AI blast radius in production
A software team's experience with a large language model upgrade highlights the challenges of managing AI blast radius in production.
VentureBeat
Jun 06, 2026·1 min read