May 2025: All AI updates from the past month

Software companies are constantly trying to add more and more AI features to their platforms, and AI companies are constantly releasing new models and features.

Here are all the major AI updates we covered in the month of May.

Mistral launches Agents API

The Agents API includes built-in connectors for code execution, web search, image generation, and MCP tools.

It offers persistent memory for conversations, allowing for seamless and contextual interactions over time. It also supports conversation branching to create new interaction paths at any point.

“The true power of our Agents API lies in its ability to orchestrate multiple agents to solve complex problems. Through dynamic orchestration, agents can be added or removed from a conversation as needed—each one contributing its unique capabilities to tackle different parts of a problem,” the company wrote in a blog post.

New Relic creates integration with GitHub Copilot’s coding agent

New Relic has announced a new integration with GitHub Copilot’s coding agent. It provides monitoring of code deployments and can automatically detect issues that arise as a result of changes.

Whenever a problem is identified, New Relic will create a GitHub issue with any relevant context, which the developer can then assign to GitHub Copilot to have it analyze the issue, draft a fix, and submit a draft pull request for the developer to review.

“With the innovative integration of New Relic’s intelligent observability technology with GitHub Copilot coding agent, we are closing the loop on ensuring continued application health,” Manav Khurana, chief product officer at New Relic. Together with our long time partner GitHub, we are providing a new, agentic way for modern software development that uses the power of agentic AI to transform the way enterprises innovate.”

Snyk launches AI agent security platform

Snyk AI Trust Platform is designed to help software development teams working with AI mitigate business risk.

It includes features such as an AI assistant that provides security intelligence recommendations, a suite of AI-powered security agents, an AI governance solution that deploys guardrails for AI development, a framework for building and maturing an AI strategy, and the ability for Snyk’s partners to integrate its capabilities into their platforms via Snyk’s MCP server.

“I’m confident that the Snyk AI Trust Platform will be a gamechanger for global organizations looking to further invest in AI-driven development,” said Danny Allan, chief technology officer, Snyk. “Autopilot didn’t replace the need for actual pilots, and in that same vein, we envision a world where AI augments developers, but never fully replaces them. No one is better positioned than Snyk to help with the near-term strategic and practical adoption of AI by building in security from the outset.”

Cast AI introduces Database Optimizer 

Database Optimizer (DBO) utilizes intelligent caching to improve cloud database performance. It works by providing an AI agent that runs a fully autonomous caching layer. This agent acts as a plug-and-play solution, with no required application changes or tuning.

“Databases are one of the most expensive components in the cloud. It was a natural addition to our Application Performance Automation platform. Our DBO agent increases performance and reduces cost by caching any queries for which it detects a repeat pattern. DBO radically simplifies caching, slashes expenses, and requires no changes to how your apps are built,” said Laurent Gil, co-founder and president of Cast AI. “APA is the future of autonomous computing —high performance, zero overhead.”

DataRobot launches syftr open source framework for agentic AI

Syftr helps developers discover and implement the best combination of components, parameters, tools, and strategies for agentic AI. Developers can use it to evaluate any module, flow, embedding model, or LLM.

“As organizations increasingly explore agentic AI systems, practitioners and developers need to quickly evaluate the latest technologies and ensure that their agentic workflows are optimally performant for specific use cases based on model quality, cost, and desired behavior. Syftr addresses this challenge through a groundbreaking multi-objective approach that rapidly simulates possible configurations to identify the best AI workflows with enterprise data and optimizes for task accuracy, latency, and cost,” the company wrote in an announcement.

Anthropic launches Claude 4 models

Claude Opus 4 and Claude Sonnet 4 are capable of undertaking long-running tasks and can work continuously for several hours. Claude Opus 4 excels at coding and complex problem-solving, whereas Claude Sonnet 4 improves on Sonnet 3.7 and balances performance and efficiency.

In addition to releasing these new models, the company also revealed a beta for extended thinking with tool use, the ability to use tools in parallel, and general availability of Claude Code.

The Anthropic API also added four new capabilities: the code execution tool, MCP connector, Files API, and the ability to cache prompts for up to one hour.

OpenAI announces software engineering agent Codex

Codex is a cloud-based agent capable of working on multiple tasks simultaneously, such as writing features, answering questions about the codebase, fixing bugs, and proposing pull requests. Each of these tasks runs separately in its own cloud sandbox environment.

“Once Codex completes a task, it commits its changes in its environment. Codex provides verifiable evidence of its actions through citations of terminal logs and test outputs, allowing you to trace each step taken during task completion. You can then review the results, request further revisions, open a GitHub pull request, or directly integrate the changes into your local environment,” OpenAI wrote in a post.

It is currently available as a research preview for ChatGPT Pro, Team, and Enterprise users, and support for Plus users will be available soon.

Amazon announces GA of AWS Transform for .NET

The new tool is an AI agent that helps with modernizing .NET applications, porting them from .NET Framework to cross-platform .NET, which, when done manually is “a labor-intensive and error-prone process. You have to perform multiple steps, such as analyzing the codebase, detecting incompatibilities, implementing fixes while porting the code, and then validating the changes. For enterprises, the challenge becomes even more complex because they might have hundreds of .NET Framework applications in their portfolio,” AWS wrote.

AWS Transform for .NET can help companies overcome these challenges. Since releasing the tool in private preview last year, Amazon has added several new capabilities, such as support for projects that have private NuGet package dependencies, support for executing unit tests once porting is complete, and the ability to port model-view-controller (MVC) Razor views to ASP .NET Core Razor views.

Anthropic launches new bug bounty program

The purpose of the program is to enable the community to stress test Anthropic’s latest safety measures. Researchers will be asked to find universal jailbreaks in safety classifiers before they are deployed publicly.

Specifically, they will test an updated version of the Constitutional Classifiers system, which is a technique developed to “guard against jailbreaks that could elicit information related to CBRN (chemical, biological, radiological, and nuclear) weapons.” It includes a list of principles defining what content should and shouldn’t be allowed when interacting with Claude.

Anthropic is teaming up with HackerOne to run the program, and it will offer up to $25,000 in rewards for jailbreaks found.

GitLab 18 integrates AI capabilities from Duo

GitLab has announced the latest version of its platform, incorporating more AI capabilities from Duo, the company’s suite of AI solutions, into the flagship DevSecOps platform.

GitLab 18 includes Duo’s AI-powered Code Suggestions for code completion and code generation, and AI-powered Chat for code explanations, code refactoring, test generation, and code fixes. Those capabilities are available for Premium and Ultimate users.

“Today’s fragmented landscape of AI point solutions creates unnecessary complexity for development teams,” said David DeSanto, chief product officer at GitLab. “By natively integrating the essential capabilities of GitLab Duo Code Suggestions and Chat directly within the GitLab DevSecOps platform, we’re eliminating the need for separate tools, licenses, and governance structures. This unified approach enables teams to accelerate their workflows and improve productivity while maintaining security and compliance standards.”

Boomi launches AI agent management solution at Boomi World 2025

At its annual conference Boomi World 2025, Boomi announced the general availability of Boomi Agentstudio (initially called Boomi AI Studio), an AI agent management solution.

With Boomi Agentstudio, developers are given a no-code platform for designing, governing, and orchestrating AI agents.

It supports the Model Context Protocol (MCP), an open standard for connecting AI agents to data sources and tools.

Boomi also announced an integration with Amazon Q Business, allowing agents created in Agentstudio to be grounded in a company’s entire knowledge base. “Imagine an agent that doesn’t just react to shipping delays based on general knowledge gleaned from the internet but instead makes intelligent decisions based on customer contracts, business rules, previous issue resolutions, partner networks, and other data — all of which is available through a company’s knowledge base,” Boomi wrote in a blog post.

CodeRabbit brings AI-powered code review into Visual Studio Code

AI code review provider CodeRabbit announced it’s bringing its solution to the Visual Studio Code editor, shifting code review left into the IDE. This integration places CodeRabbit directly into the Cursor code editor and Windsurf, the AI coding assistant purchased recently by OpenAI for US$3 billion.

By bringing CodeRabbit into VS Code, Cursor, and Windsurf, CodeRabbit is embedding AI at the earliest stages of development. “As we are bringing the reviews within the editor, then those code changes could be reviewed before each are pushed to the central repositories as a PR and also before they even get committed, so that developer can trigger the reviews locally at any time,” Gur Singh, co-founder of the 2-year-old CodeRabbit, told SD Times.

Zencoder launches Zen Agents

Zen Agents are customizable AI agents that understand your code and connect with your tools. Zencoder has launched a platform for customizing them, as well as an open-source marketplace where the community can contribute their own agents for others to use.

“Zen Agents create the perfect harmony between human creativity and targeted AI assistance,” said Andrew Filev, CEO and founder of Zencoder. “By enabling teams to craft agents with specific expertise and then deploy them organization-wide, we’re helping developers achieve that elusive technical flow state where complex problems seem to solve themselves.”

Anaconda launches unified AI platform

Anaconda AI Platform is a platform that brings together all of the tools needed to source, secure, build, and deploy AI in open source ecosystems.

It offers Quick Start Environments that are pre-configured, vetted for security, and tailored for Python, finance, and AI/ML development. According to Anaconda, its platform can reduce or eliminate the need to manage configurations so that developers have more time to work on building.

“The Anaconda AI Platform addresses the evolving needs of our ever-growing user base,” said Laura Sellers, co-president and chief product and technology officer at Anaconda. “Last year, Anaconda customers quadrupled to over one million, underscoring the need for more efficient, secure, and integrated package security management for AI innovation with open source. We are focused on giving everyone the confidence and clarity they need to accomplish their data science and AI goals.”

Parasoft adds agentic AI capabilities to SOAtest

SOAtest’s AI Assistant now uses agentic AI to help with test scenario generation, such as generating test data and parameterizing test scenarios for data looping.

It can work through multi-step workflows alongside the tester, allowing testers to execute comprehensive tests without scripts, advanced code-level skills, or in-depth domain knowledge.

IBM introduces new tools to help with scaling AI agents across the enterprise

At its IBM THINK conference earlier this week, IBM introduced new updates that will help alleviate some of the challenges associated with scaling AI agents.

New agent capabilities in watsonx Orchestrate include:

• New tools for integrating, customizing, and deploying agents
• Pre-build domain agents for HR, sales, and procurement
• Integration with over 80 enterprise applications, including ones from Adobe, AWS, Microsoft, Oracle, Salesforce Agentforce, SAP, ServiceNow, and Workday
• Agent orchestration capabilities for complex projects like workflow planning and task routing that require coordination between multiple agents and tools
• Agent observability across the entire agent life cycle

The company also announced its Agent Catalog to provide easier access to agents from IBM and its partners.

Anthropic adds web search capabilities to its API

This latest addition will enable developers to build applications and agents that can access and deliver the most up-to-date insights.

“When Claude receives a request that would benefit from up-to-date information or specialized knowledge, it uses its reasoning capabilities to determine whether the web search tool would help provide a more accurate response. If searching the web would be beneficial, Claude generates a targeted search query, retrieves relevant results, analyzes them for key information, and provides a comprehensive answer with citations back to the source material,” Anthropic wrote in a blog post.

Amazon Q Developer gets new agentic coding experience in Visual Studio Code

Amazon has announced a new agentic coding experience for Amazon Q Developer in Visual Studio Code.

“This experience brings interactive coding capabilities, building upon existing prompt-based features. You now have a natural, real-time collaborative partner working alongside you while writing code, creating documentation, running tests, and reviewing changes,” Amazon wrote in a blog post announcing the news.

Google releases updated version of Gemini 2.5 Pro Preview

The updates implement better coding capabilities, especially for tasks like transforming code and creating agentic workflows.

According to Google, this release addresses developer feedback such as reducing errors in function calling and improving function calling trigger rates.

OpenAI to buy Windsurf

Bloomberg reported the deal earlier this week, saying that OpenAI would acquire the company for $3 billion. According to Bloomberg, the deal has not yet closed.

Windsurf, previously called Codeium, is an agentic IDE designed to enable seamless collaboration between developers and AI.

HCL announces new AI agent orchestration platform

HCL Universal Orchestrator (UnO) Agentic is an orchestration platform for coordinating workflows among AI agents, robots, systems, and humans.

It builds upon HCL’s Universal Orchestrator, and adds agentic AI capabilities to provide intelligent orchestration and insert AI agents into business-critical processes and workflows.

“By integrating deterministic and probabilistic execution, HCL UnO transforms how humans and intelligent systems collaborate to shape the future of enterprise operations,” said Kalyan Kumar (KK), chief product officer of HCLSoftware.

DigitalOcean announces new NVIDIA-powered GPU Droplets

NVIDIA RTX 4000 Ada Generation, NVIDIA RTX 6000 Ada Generation, and NVIDIA L40S GPUs are now available as GPU Droplets.

According to Bratin Saha, chief product and technology officer at DigitalOcean, the new offerings are meant to provide customers with access to more affordable GPUs for their AI workloads.

“DigitalOcean’s simple and scalable cloud platform makes it easier to deploy advanced AI workloads on NVIDIA technology, so organizations can quickly and more easily build, scale, and deploy AI solutions,” said Dave Salvator, director of accelerated computing products at NVIDIA.

Yellowfin 9.15 now available

The latest version of the business intelligence platform introduces AI-enabled Natural Query Language (AI NLQ), which allows users to ask questions about their data.

Other updates in this release include expanded REST API capabilities, enhanced bar and column chart customization, simpler yearly data comparisons and report styling, stricter default controls for better data security, and support for writable Clickhouse data sources.

“Yellowfin 9.15 debuts the first integration between the Yellowfin product and AI platforms,” said Brad Scarff, CTO of Yellowfin. “These platforms have enormous potential to unlock productivity and usability benefits for all of our customers, and upcoming versions of Yellowfin will build on this initial release to provide further innovative AI-enabled features.”

Apiiro announces partnership with ServiceNow

As a result of the collaboration, Apiiro’s AI-native deep code analysis (DCA) and code-to-runtime matching will be utilized in ServiceNow’s Configuration Management Database (CMDB), which provides an up-to-date view of IT and software environments

“This integration is a major milestone for Apiiro and the ASPM market at large, as IT operations, security operations, and application security continue to converge,” said John Leon, VP  of partnerships and business development at Apiiro. “It’s a privilege to expand our partnership with ServiceNow by introducing our Agentic Application Security platform as the definitive source of truth for software development and becoming the software development lifecycle (SDLC) Systems of Record within the ServiceNow CMDB, equipping enterprise users with a precise inventory of software assets to ensure operational efficiency in today’s rapidly evolving, AI-driven software development revolution.”

Dremio launches MCP Server

The server will allow AI agents to explore datasets, generate queries, and retrieve governed data.

“Dremio’s implementation of MCP enables Claude to extend its reasoning capabilities directly to an organization’s data assets, unlocking new possibilities for AI-powered insights while maintaining enterprise governance,” said Mahesh Murag, product manager at Anthropic.