Anthropic has officially launched Claude Sonnet 5, a new “agentic” AI model designed for complex task automation, coding, and tool use. The release comes as the company navigates a shifting regulatory landscape, with the U.S. Department of Commerce confirming the lifting of export controls on the company’s more powerful Fable 5 and Mythos 5 models.
The Capabilities of Sonnet 5
Claude Sonnet 5 is positioned as the most agentic model in the Sonnet lineup, capable of autonomous planning and tool execution. According to Anthropic, it outperforms its predecessor, Sonnet 4.6, in reasoning, coding, and technical grounding. The model is currently available across all subscription tiers, including Free, Pro, and Enterprise, with introductory pricing set at $2 per million input tokens until August 31, 2026.
Early feedback from partners, including firms like Lovable and ClickHouse, suggests that Sonnet 5 excels in “brownfield” code environments—handling complex debugging and multi-step workflows that previously stalled smaller models. Despite its increased capabilities, Anthropic reports that Sonnet 5 maintains a lower rate of undesirable behaviors compared to previous iterations and includes built-in cyber safeguards by default.
Regulatory Detente
The launch coincides with a significant policy shift in Washington. On June 30, 2026, the U.S. Department of Commerce officially lifted export controls that had forced Anthropic to pull the Fable 5 and Mythos 5 models offline on June 12. The ban was originally enacted following concerns that users could circumvent safety guardrails, posing a potential national security risk.
Commerce Secretary Howard Lutnick cited Anthropic’s “close coordination and cooperation” with federal officials as the primary driver for the decision to restore access. Anthropic has since implemented specific safeguards to address vulnerabilities identified by partners like Amazon. Industry observers note that this resolution highlights the ongoing tension between rapid AI development and the government’s desire for oversight, with firms like OpenAI expressing caution regarding the precedent of government-mandated release phases.
Market and Strategic Implications
The return of Fable 5 and the introduction of Sonnet 5 reflect a broader strategy of balancing power with safety. While Mythos 5 remains restricted to select government-approved entities due to its high-speed vulnerability exploitation capabilities, the widespread release of Sonnet 5 provides developers with a high-performance, cost-effective alternative for production environments. As Anthropic continues to work within the framework of the Cyber Verification Program, the company’s ability to maintain its competitive edge while satisfying federal security mandates will likely serve as a litmus test for the wider AI industry.