Post by @f9finance

Anthropic launched Claude Fable 5 today, the first Mythos-class model, a tier above Opus. I tested it on real FP&A work the same day: cross-file data modeling, a controller-grade review pass, and a full treasury analysis from one prompt. This isn't a benchmark review. It's a stress test of what AI for finance looks like when the underlying model can actually hold long, multi-file tasks together, the kind of work where Claude for finance lives or dies on real data, not synthetic prompts. You'll see Claude Fable 5 join 149,000 POS transactions to GL data with no common key, cross-check variance commentary against actuals, and run a 5-step treasury review without a single follow-up prompt. I also cover what Fable 5 costs, the free access window through June 22, the Opus 4.8 fallback behavior, and when Sonnet 4.6 is still the smarter pick for everyday FP&A with AI. If you work in FP&A, accounting, or finance automation, this is what the new tier actually means for your workload, tested on real files instead of benchmarks. 🛠️ Tools Used - Claude Fable 5 (Anthropic's new Mythos-class model) - Claude desktop app - Claude Opus 4.8 / Sonnet 4.6 (tier comparison) - Microsoft Excel — real P&L, GL, and POS dataset