Anthropic Launches Claude Fable5.1 and Mythos5.1: Performance Gains and Price Cuts
APPSO · wechat · 2026-09-02
Anthropic released Fable5.1 and Mythos5.1, sharing the same base model but differing in safety permissions. Fable5.1 is general-purpose, while Mythos5.1 offers relaxed limits for cybersecurity and bio research via a trusted access program.
Key Upgrades
- Performance: Achieves better results than Fable5 at lower cost in Low/Medium modes. Real-world tests show strong gains in code generation, long-haul agent tasks, and research (e.g., protein design, GPU kernel optimization).
- Autonomous Capability: Supports multi-hour unattended runs, including a 38-hour autonomous ML experiment and solving a 370-year-old cipher.
Pricing
- Cache Reduction: Cache Read costs dropped by 75% to $0.25/M tokens. Overall costs drop by 25% for typical workloads and up to 45% for complex agents.
- Competitive Landscape: Despite cuts, Fable5.1 remains 6x more expensive than GPT-5.6 Sol for similar tasks, offering richer details but still prone to state errors.
Safety & Compliance
- Enterprise Safeguards (EFS): New system allowing enterprise data to reside in customer-controlled clouds, addressing compliance concerns (rolling out in Autumn).
- Refined Safety: Cybersecurity interventions reduced by 60%; bio-science false triggers reduced by 85%. Allows more defensive vulnerability analysis.
- Anti-Distillation & Watermarking: New API keys cannot modify history context to prevent distillation. Digital watermarks added for EU compliance, with a detection API in preview.
Industry Insight
The article discusses the "Intelligence Density Kill Line," where competitors like Zhipu and Alibaba offer near-flagship performance at a fraction of the cost, challenging the "Leader's Curse" of sustaining high R&D and infrastructure costs for marginal gains.
More from Models
- Shopify's 0.8B fine-tune beats GPT 5.6, saves millions — CShorten30 · 2026-09-02
- Rumor: Gemini 3.8 Flash Arrives Tomorrow, Skipping 3.5 Pro — Hesamation · 2026-09-02
- Claude Fable 5.1 Replicates and Extends Research; Mythos Writes Custom Kernels — BenBlaiszik · 2026-09-02
- Developer finds flaws in CritPt benchmark scores — scaling01 · 2026-09-02
- MiniMax-M3 hits 2,274 tok/s in Cacheon arena, focused on enterprise end-to-end testing — const_reborn · 2026-09-02
- Claude flags garden keep-out zones, suggests reorienting camera for safety — _Stocko_ · 2026-09-02