Anthropic and OpenAI target big businesses with enterprise-grade controls and lower pricing - SiliconANGLE

Anthropic and OpenAI target big businesses with enterprise-grade controls and lower pricing

Artificial intelligence leaders Anthropic PBC and OpenAI Group PBC are stepping up their efforts to compete for the enterprise, making their most advanced agentic tools more accessible to the largest organizations.

In its update today, Anthropic revealed new "organization-wide controls" to help corporate teams deploy its autonomous Claude Cowork service. Meanwhile, OpenAI took a different path, slashing the cost of the "Pro" subscription to access its popular Codex programming tools.

Anthropic said Claude Cowork is getting a variety of new administrative tools designed to help organizations manage its rollout. Claude Cowork, first announced in January, is an autonomous AI agent that's able to handle complex, multistep tasks on employee's computers. Unlike standard chatbots that just answer questions, it's more like a proactive teammate, capable of organizing files, creating reports and documents and running browser tasks, rather than just talking about them.

While Claude Cowork is every bit as good as Claude Code at writing software, most early adopters can be found in non-engineering departments such as marketing, financial and legal teams, where they're using it to help with project updates and research tasks.

To facilitate this reality, Anthropic has introduced role-based access controls for Enterprise subscribers, which enable admins to choose exactly which Cowork capabilities employees can access. There are also new group spend limits to support per-team budgeting, delivering the cost predictability that financial teams need.

On the technical side, the company is enhancing observability by expanding Cowork's OpenTelemetry support. This means companies can now monitor Claude events such as tool calls and file modifications directly within their security information and event management pipelines. In addition, the company announced a new Zoom Model Context Protocol connector that allows Claude Cowork to pull meeting summaries and action items into its workflows, and more precise controls for other MCP connectors.

Not to be outdone, OpenAI is pushing to entice the heaviest enterprise users to engage with Codex, its rival to Claude Code. The company is lowering the barrier to entry with the availability of a new $100 per month Pro plan, effectively slashing the cost of its most premium subscription in half.

The new tier is aimed at Codex power users, offering five times as much usage as the existing $20 pricing tier. It's aimed at professional developers who keep finding themselves hitting limits during intensive coding sessions, and it dramatically undercuts the cost of rival coding services. For instance, the top-tier subscriptions of Claude Code and Google LLC's Gemini Code Assist both start at $200 per month.

In addition to the lower costs, Pro Codex users will also be able to access experimental preview features enabled by OpenAI's latest frontier models, including GPT-5.4 Pro.

Today's updates underline the strategic shift as AI's top model makers scramble to sign up more enterprises in the belief that businesses will ultimately become a major source of revenue in the years to come. They're fighting tooth and nail to win over customers, and that means they cannot only compete on model performance, but also areas such as utility and integration.

The biggest beneficiaries are enterprises themselves. With Anthropic helping to solve the logistical headaches and OpenAI perhaps kicking off a deflationary trend in AI compute, it becomes easier for big businesses to adopt the most powerful AI agents and start accelerating automation.