Anthropic has unveiled Claude Fable 5.1 and Mythos 5.1, with Fable 5.1 doubling scores on key science benchmarks and slashing cache read fees by 75%. The models are now available across major cloud platforms.
TL;DR
- Claude Fable 5.1 scores 52.6% on Terminal-Bench-Science 0.1, more than double its predecessor's score.
- Cache read fees for Fable 5.1 reduced by 75% to $0.25 per million tokens.
- Anthropic introduces Enterprise Frontier Safeguards for enhanced data security.
What happened
Anthropic launched Claude Fable 5.1 and Mythos 5.1 on September 1, positioning them as advanced models for coding and knowledge work. Fable 5.1 is available to all users, while Mythos 5.1 is restricted to a trusted access program for high-risk sectors like cybersecurity and life sciences.
Fable 5.1 outperformed its predecessor and competitors on multiple benchmarks. It scored 52.6% on Terminal-Bench-Science 0.1, more than double Fable 5's score, and surpassed Opus 5 and GPT-5.6 Sol. On Terminal-Bench 4.0 for coding, Fable 5.1 scored 55.8%, with Mythos 5.1 reaching 60.9%.
Anthropic reduced cache read fees for Fable 5.1 by 75% to $0.25 per million tokens, maintaining base pricing at $10 for input and $50 for output. The company claims overall cost reductions of up to 45% under highly agentic workloads.
Why it matters
For developers and startups, the improved performance and reduced costs of Fable 5.1 offer a more powerful and affordable tool for coding and knowledge work. The tiered access model allows for safer deployment in high-risk scenarios.
Investors should note the competitive edge Anthropic gains through these advancements, particularly in benchmark performance and cost efficiency. The introduction of Enterprise Frontier Safeguards also enhances data security, a critical factor for enterprise adoption.
The competitive landscape is reshaped by Fable 5.1's superior benchmark scores, challenging OpenAI's GPT-5.6 Sol and Anthropic's own Opus 5. The reduced fees and improved safety mechanisms make it an attractive option for both individual and enterprise users.
Key facts
- Claude Fable 5.1 scores 52.6% on Terminal-Bench-Science 0.1, more than double Fable 5's score of 26.3%.
- Fable 5.1 scores 55.8% on Terminal-Bench 4.0, with Mythos 5.1 reaching 60.9%.
- Cache read fees for Fable 5.1 reduced by 75% to $0.25 per million tokens.
- Base pricing remains unchanged at $10 per million tokens for input and $50 per million tokens for output.
- Enterprise Frontier Safeguards (EFS) allows customers to store data in their own controlled cloud infrastructure.
- Fable 5.1 reduces cybersecurity false-positive blocks by approximately 60%.
- Models available across Claude API, Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry.
- Fable 5.1 will be retired no earlier than September 1, 2027.
Context
Anthropic's release of Claude Fable 5.1 and Mythos 5.1 comes at a time of intense competition in the AI model space. With advancements in benchmark performance and cost efficiency, Anthropic is positioning itself as a strong contender against OpenAI and other major players.
The introduction of Enterprise Frontier Safeguards reflects a growing emphasis on data security and privacy in the AI industry. This move is likely to attract enterprise customers who prioritize data control and compliance.
The tiered access model for Mythos 5.1 highlights the need for balanced capability deployment in high-risk sectors. This approach allows for the safe and responsible use of advanced AI models in sensitive areas like cybersecurity and life sciences.
