Anthropic has launched Claude Fable 5.1 and Mythos 5.1, positioning them as its "world’s most advanced models for coding and knowledge work." Despite unchanged base prices, customers are seeing significant cost reductions.

• The new models are approximately 25% cheaper to run for most customers due to aggressive cost reductions in cached token processing.
• Anthropic has refined safety filters, with cybersecurity filters firing 60% less often in Claude Code and biology filters 85% less often on ordinary medical questions.
• Mythos 5.1 offers loosened safety filters specifically for vetted cybersecurity defenders and life scientists, currently available to select US organizations.
• Performance benchmarks show substantial improvements, with Fable 5.1 scoring 52.6% on Terminal-Bench-Science and 55.8% on Terminal-Bench for coding tasks.