Anthropic has officially launched its newest artificial intelligence models, Claude Fable 5.1 and Mythos 5.1. The release of these systems arrives at a critical moment for the industry, as AI agent security incidents are currently drawing increased scrutiny.

In response to this climate, Anthropic has heavily touted the safety gains incorporated into the new models.

The newly introduced systems were specifically developed to address a series of prominent criticisms from the company's customer base. According to Anthropic, the deployment of Fable 5.1 and Mythos 5.1 directly tackles ongoing user concerns regarding high prices, data retention policies, and historically overzealous safety safeguards.

Cost reductions and performance upgrades

The company claims that Claude Fable 5.1 delivers stronger overall performance compared to its predecessor, Fable 5. Despite this upgrade in underlying capabilities, the new model typically costs around 25 percent less for standard use cases.

For users running more advanced workloads, the financial savings can be even more substantial. Anthropic states that Claude Fable 5.1 is up to 45 percent cheaper for complex agentic tasks. This significant cost reduction is achieved thanks to reduced pricing on cached data that the system has already processed and stored.

In addition to the base price drops, Anthropic has made structural adjustments to how the model handles user prompts. Fable 5.1 includes specific changes meant to reduce overall token costs for developers.

Furthermore, the update directly addresses prior complaints about moderation by refining the model's behavioral guardrails. Anthropic implemented changes specifically designed to reduce false-positive restrictions that were previously triggered by the model's safety safeguards.