Anthropic launches Claude Fable 5.1, cuts costs up to 45%

The launch of Claude Fable 5.1 directly addresses the primary criticism leveled at its predecessor. The article notes that high cost was the biggest complaint about Fable 5 and likely contributed to low adoption among enterprise customers. By making the model cheaper, especially for the heavy agentic workloads that represent the frontier of AI use, Anthropic is attempting to remove the financial barrier that has hindered its enterprise growth. The price pressure had been building since Opus 5 launched in late July, already matching or beating Fable 5 on most benchmarks at a lower price.

The benchmark results illustrate the model's technical progress, particularly in agentic tasks. On Terminal-Bench 4.0 for agentic coding, Fable 5.1 scores 55.8 percent, up from 42.0 percent for Fable 5 and 37.3 percent for GPT-5.6 Sol. The model also improved its writing style, reducing reliance on bullet points and bold text in chat. However, whether these gains translate to real-world use at the same scale will become clear over the coming weeks, as the article notes.

The introduction of built-in watermarks and a detection API marks a shift toward accountability. This feature is aimed at regulators, media outlets, fact-checkers, and research institutions, allowing them to verify if text contains a watermark. The company also addressed safety filters, which had been over-sensitive in 5.0 models, and has taken steps to close a documented distillation attack technique. These moves suggest Anthropic is focusing on both capability and responsible deployment as it expands access to its most advanced systems.