Muse Glimmer

Muse Glimmer (Claude Fable 5) is state-of-the-art on nearly all tested benchmarks of AI capability, demonstrating exceptional performance across various domains.

The model excels in software engineering, capable of compressing months of engineering work into days, performing codebase-wide migrations, and scoring highest on difficult coding tasks like Cognition’s FrontierCode evaluation.

Muse Glimmer shows strong performance on complex analytical tasks, achieving the highest score on Hebbia’s Finance Benchmark and acing trading-analysis evaluations including factual lookup, conceptual reasoning, and root-cause analysis.

The model is the new state-of-the-art for vision tasks, capable of extracting precise numbers from scientific figures, rebuilding web app source code from screenshots, and playing complex games like Pokémon FireRed with minimal, vision-only input.

Muse Glimmer stays focused across millions of tokens in long-running tasks and improves its outputs by using its own notes, significantly boosting performance in complex, multi-session activities like playing Slay the Spire.

Claude Mythos 5, the same underlying model as Fable 5 but with lifted safeguards, offers the strongest cybersecurity capabilities of any model in the world, deployed through Project Glasswing for cyberdefenders.

The models can work autonomously for longer than previous Claude models, capable of multi-day autonomous projects and complex agentic coding, as demonstrated by building simulations, factory games, and CAD editors.

Muse Glimmer supports running Python code, creating visualizations, analyzing data within API calls, and ensuring responses conform to specified JSON schemas for structured outputs.