Anthropic has extended the promotional period for free use of the Claude Fable 5 model for a second time until July 19, 2026. Simultaneously, a market shift toward ultra-efficient autonomous agents is emerging with the launch of Oak Lab.
What Happened
Anthropic has provided users on Pro, Max, and Team plans the ability to use up to 50% of their weekly limit on the Claude Fable 5 model without additional charges. Additionally, a bonus is in effect for the Claude Code tool, providing a 50% increase in quotas. In parallel, Richard Sutton announced the launch of the startup Oak Lab, which plans to develop autonomous agents based on continual reinforcement learning (continual RL), targeting 1 trillion parameters with power consumption of approximately 20W.
Context
The AI market is demonstrating a transition from static models to dynamic agentic systems. While Anthropic is lowering the barriers to entry for testing agentic capabilities via Claude Code, projects like Oak Lab aim to solve the problems of energy efficiency and the ability of models to learn in real-time, rather than solely on fixed datasets.
Why It Matters for the Industry
For the industry, this means an acceleration in the adoption of agentic technologies: developers can scale the testing of complex scenarios in Claude Code for free. The emergence of Oak Lab signals a potential shift in architectural standards, where the focus moves toward extreme energy efficiency and the dynamic environmental cognition of agents.
Why It Matters for Users
Regular users and developers can test the capabilities of the top-tier Claude Fable 5 model for free until July 19 using their current subscriptions. It is also worth watching the development of Oak Lab's technologies, as their approach to continual learning could radically reduce the Total Cost of Ownership (TCO) for future autonomous systems.
What Is Not Yet Known / Limitations
Anthropic's promotion is limited by the lack of API access during the free period, which complicates industrial implementation into production pipelines. For the corporate sector and legal professionals, use of the model may be limited by security and liability concerns.
Sources
Author
Look at AI, Editorial Team