Parameters
Undisclosed
Context Length
1M
Modality
Multimodal
Architecture
Undisclosed
License
Proprietary
Release Date
1 Sept 2026
Knowledge Cutoff
Jun 2026
API Pricing (per 1M)
Input: $10.00 · Output: $50.00
Rank
#1
| Benchmark | Score | Rank |
|---|---|---|
General | max 0.83 | 🥇 1 |
Coding | max 0.86 | 🥇 1 |
Mathematics | max 0.97 | 🥇 1 |
LiveBench Average | max 0.83 | 🥇 1 |
Agent Arena | max 0.14 | 🥇 1 |
Intelligence Index | max 0.53 high 0.51 medium 0.49 low 0.47 | 🥇 1 🥉 3 8 12 |
Coding Index | max 0.82 high 0.79 medium 0.77 low 0.75 | 🥇 1 🥈 2 6 22 |
Agentic Index | max 0.58 high 0.53 medium 0.51 low 0.48 | 🥇 1 8 15 22 |
Reasoning | max 0.92 | 🥈 2 |
Agentic Coding | max 0.66 | 🥈 2 |
Web Development | max 1758 | 🥈 2 |
Data Analysis | max 0.80 | 4 |
General Text | max 1498 | 5 |
Overall Rank
#1 🥇
Coding Rank
#1 🥇
Claude Fable 5.1 is an advanced multimodal reasoning and agentic model by Anthropic, offering improvements across coding, long-running agent workflows, and knowledge tasks. It features a 1M token context window tailored for extensive codebase refactoring and visual analysis.
Architecture specifications are undisclosed for proprietary models.
Anthropic's Claude Fable 5 generation, released June 9, 2026, shares a unified frontier architecture with Mythos 5, incorporating advanced safety classifiers for cybersecurity, biology, and chemistry, with fallback routing to Claude Opus 4.8.
Assistant
Online