ADB Target Private Investment for Broader Economic Growth
September 4, 2026Airlift Founder’s Metal Raises $4.5M to Reinvent Fundraising
September 4, 2026
Anthropic has launched Claude Fable 5.1 and Mythos 5.1, its most advanced models for coding and scientific research, at 25% lower cost and with significantly improved safeguards.
Anthropic has launched what it describes as the world’s most advanced AI models for coding and knowledge work. Claude Fable 5.1 and Claude Mythos 5.1, released in September 2026, set new benchmarks across agentic coding, scientific research, multidisciplinary reasoning, and business workflow automation, while arriving at a meaningfully lower cost than their predecessors.
Fable 5.1 and Mythos 5.1 share the same underlying model but differ in their safeguard configurations. Fable 5.1 is generally available, while Mythos 5.1 is accessible only through Anthropic’s trusted access programmes, specifically designed for vetted professionals in cybersecurity and the life sciences.
The performance gains are substantial. On Terminal-Bench-Science 0.1, Fable 5.1 scored 52.6%, nearly doubling Fable 5’s 24.7%. On CursorBench 3.2.0, it achieved 73.4%, ahead of both Fable 5 at 70.5% and OpenAI’s GPT-5.6 at 67.2%.
On Humanity’s Last Exam, it reached 65.0% with tools. Early access partners from Jane Street Capital, MongoDB, Shopify, Canva, Square, Plaid, and Ramp all reported significant qualitative improvements across coding, research, and writing tasks.
The pricing shift is equally notable. Cache reads now cost 75% less at $0.25 per million tokens, translating to approximately 25% lower costs for typical workloads and up to 45% for highly agentic tasks. Fable 5.1’s input and output token pricing remains unchanged at $10 and $50 per million tokens respectively.
On the scientific frontier, Mythos 5.1’s capabilities are extraordinary. In molecular design, it produced protein binders with affinities 10 times higher than the best designs submitted to Adaptyv Bio’s protein design competitions, achieving a hit rate of nearly 50% across 12 targets, compared to the industry standard of 10 to 15%.
Fable 5.1 separately trained a neural network to produce a new high-resolution elevation map of one third of Venus using NASA Magellan radar data from more than 30 years ago, revealing surface details down to two to three kilometres rather than the previous 10 to 20.
Anthropic has released this map under a Creative Commons licence ahead of upcoming NASA VERITAS and ESA EnVision missions.
On the safety front, Anthropic introduced Enterprise Frontier Safeguards (EFS), a new system that provides enterprise customers complete data privacy equivalent to zero data retention, while storing data on customer-controlled cloud infrastructure rather than Anthropic’s own systems.
Cybersecurity safeguards have also been refined to block 60% fewer false positives per session than before, while still restricting exploit generation and penetration testing tasks. Alignment evaluations showed Mythos 5.1 is less likely than its predecessor to attempt resource access outside its test environment, use motivated reasoning to justify actions, or succeed at reward hacking.
Fable 5.1 is available today on Amazon Web Services, Google Cloud, and Microsoft Azure via the model string claude-fable-5-1.
