Anthropic launches Claude Fable 5.1 with cheaper cache reads for agentic workloads

Anthropic has released **Claude Fable 5.1**, a new generally available model for coding and knowledge work, alongside the more restricted **Claude Mythos 5.1**. The release changes both capability positioning and the economics of long-running Claude workloads.\n\nFable 5.1 remains priced at $10 per million input tokens and $50 per million output tokens, but Anthropic has cut cache reads to $0.25 per million tokens — 75% below Fable 5. Anthropic estimates that this reduces typical billed workloads by about 25% and highly agentic workloads by as much as roughly 45%. Those savings are vendor estimates, not an independent benchmark of every workload.\n\n## The pricing change targets agents, not headline token rates\n\nThe base input and output rates did not fall. The meaningful change is the lower cost of reusing previously processed context, which matters most when an agent repeatedly consults a large working set over long tasks. Independent launch coverage similarly highlights cache pricing as the core economic change rather than a broad reduction in list price.\n\nAnthropic says Fable 5.1 is available to Pro, Max, Team and Enterprise users and through the Claude Platform, major cloud marketplaces and partner clouds. Mythos 5.1 uses the same underlying model with different safeguards and remains limited to vetted cybersecurity and life-sciences users.\n\n## What AiToolMap should update\n\nThe release materially changes Claude’s current model lineup, pricing details and safety/access segmentation. AiToolMap should therefore treat it as a review-update trigger for Claude, while keeping benchmark and savings claims clearly attributed to Anthropic until independent testing accumulates.