AI Headlines: OpenAI Codex Adds Internet Access, Mistral Releases Coding Assistant

Software companies are constantly trying to add more and more AI features to their platforms, while AI companies are continuously releasing new models and capabilities. Keeping up with all of this can be difficult, so we have written this summary to share several important AI updates that developers should be aware of.

OpenAI Announces Codex Update

The coding agent Codex now has internet access during task execution, enabling new capabilities such as installing runtime basic dependencies, testing required external resources, and upgrading or installing packages.

Currently, internet access is turned off by default. You can enable it when creating a new environment or edit an existing environment to turn it on. Users can control the domains and HTTP methods that Codex can use.

OpenAI also announced that Codex has begun rolling out to ChatGPT Plus users. The company also noted that during peak demand periods, Plus users may experience rate limits.

Mistral Releases New Coding Assistant

Mistral Code was recently released, based on an open-source project that provides a central hub for models, rules, prompts, documentation, and other building blocks for creating AI code assistants. It is powered by four different coding models: Codestral, Codestral Embed, Devstral, and Mistral Medium.

It is now proficient in over 80 programming languages and can reason about files, Git diffs, terminal output, and issues. Currently, an internal beta version is available in mainstream editors like JetBrains IDE and VSCode.

Mistral Code's goal is simple: to provide enterprise developers with best-in-class coding models through an integrated platform deployable in the cloud, with reserved capacity, or on isolated local GPUs, enabling functionalities from instant completion to multi-step refactoring. Unlike typical SaaS copilots, all parts of the Mistral Code stack—from models to code—are provided by the same vendor, adhere to the same SLAs, and every line of code resides within the customer's enterprise boundaries, the company wrote in its announcement.

Postman Introduces Agent Mode, Integrating AI Agent Power into Postman's Core Capabilities

Agent mode allows for creating, organizing, and updating collections; creating test cases; generating documentation; building multi-step agents to automate repeatable API tasks; and setting up monitoring and observability.

Abhinav Asthana, CEO and co-founder of Postman, mentioned that it's like having a Postman expert user by your side.

The company also announced that users can convert any public API on the Postman network into an MCP server. Additionally, the company launched an MCP server network where publishers can host agent tools and developers can easily find them. "We've essentially consolidated all currently available remote MCP servers, validated them, and then deployed them to the public network," said Abhinav Asthana, CEO and co-founder of Postman.

FinOps Foundation Launches FinOps for AI Certification

The foundation explained that the training and certification are designed to "help FinOps practitioners understand, manage, and optimize cloud computing expenditures related to artificial intelligence."

It will discuss topics such as AI-specific cost allocation, chargeback models, workload optimization, unit economics, and sustainability.

The training course will be divided into four parts, with the first part now available and the remaining parts launching in September 2025, November 2025, and January 2026. The certification exam will begin in March 2026.

Microsoft Dev Proxy Latest Version Adds LLM Usage and Cost Tracking Features

Dev Proxy 0.28 currently includes OpenAITelemetryPlugin, which allows you to understand how your application interacts with OpenAI. For each request, it will provide information on the model used, token count, cost estimation, and a grouped summary per model.

Dev Proxy can now also use the local AI runtime stack Foundry Local as its local language model provider.

Other updates in Dev Proxy 0.28 include new extensions for .NET Aspire, improved TypeSpec generation of PATCH operations, support for JSONC in mock files, and improved logging.

Snowflake Unveils Innovative Agent AI Technology to Empower Data Insights

Snowflake Intelligence (soon to be in public preview), powered by intelligent data agents, offers a natural language query experience to derive actionable insights from structured and unstructured data. According to the company's media announcement, a new data science agent is also coming soon in private preview, designed to help data scientists automate routine machine learning model development tasks.

Snowflake Intelligence brings together data from numerous sources and uses the new Snowflake Openflow to simultaneously compile information from spreadsheets, documents, images, and databases.

Data agents can generate visualizations and assist users in acting on insights, Snowflake stated in its announcement. Snowflake Intelligence can also access third-party knowledge via Cortex Knowledge Extensions, which will be generally available soon on the Snowflake Marketplace.

Progress Adds New AI Code Assistant

Progress Software announced that new AI code assistants and other features for modern application development will be built into the Q2 2025 releases of Progress Telerik and Progress Kendo UI, .NET, and JavaScript UI libraries. The company also announced that this release introduces AI coding assistants for Blazor and React, AI-powered theme generation, and GenAI-supported reporting insights.

AI coding assistants can automatically generate numerous AI IDE code for Blazor and KendoReact libraries within Telerik UI, reducing manual editing time and shortening development cycles. Additionally, the company wrote in its announcement, developers can add natural language prompts to Progress ThemeBuilder to create custom styles for Telerik and Kendo UI components. This release also includes report summaries and insights powered by generative AI in Progress Telerik Reporting, as well as a GenAI-backed PDF processing library. The company stated that this library can provide "instant document insights, AI prompt options in editor controls, and new AI building blocks and page templates to accelerate UI development."

IBM Announces Watsonx AI Lab Establishment

Watsonx AI Lab, an IBM innovation center located in New York City, will connect AI developers with IBM's resources and expertise.

IBM stated that New York City was chosen as the incubation base due to its more than 2,000 AI startups. The company hopes to support these startups and collaborate with local universities and research institutions.

"This is not your typical corporate lab. The Watsonx AI Lab gives the best AI developers access to world-class engineers and resources to build new businesses and applications that will reshape enterprise AI. We are rooting this mission in New York City to invest in a diverse, world-class talent pool and a vibrant community whose innovations have long impacted the tech landscape," said Ritika Gunnar, General Manager of Data and AI at IBM.

Editor: Field Manager

AI Headlines: OpenAI Codex Adds Internet Access, Mistral Releases Coding Assistant

Share Short URL