Claude’s new models: what Opus 4.8 and Fable 5 change for automation

Anthropic’s Claude family has grown. Two models stand out: Claude Opus 4.8, the flagship for everyday work, and Claude Fable 5, the most capable model for the hardest, long-horizon tasks. For us these aren’t a technical curiosity — they directly set what automation can reliably do.

The current lineup

Prices are per million input/output tokens. Opus 4.8 and Fable 5 both offer a 1M-token context window with no long-context premium.

Opus 4.8 — the everyday workhorse

Opus 4.8 is the most capable model in the Opus line: highly autonomous and strong at long-horizon agentic work, knowledge work, and memory-heavy tasks. In practice that means an automation can carry more steps on its own, in the right order, from a single instruction — without bailing back to a human midway. The 1M context means you can hand the model the rules of an entire operation in one shot.

Fable 5 — for the hardest work

Fable 5 is Anthropic’s most capable widely released model, for the most demanding reasoning and long-horizon agentic work. A single request can run for minutes on hard tasks — its reasoning is always on. It costs more than Opus, so we reserve it for multi-step, high-value automations where the margin for error is thin, not for routine work.

What it means for your business

The most capable model isn’t always the right answer. For most automation steps Opus 4.8 (or Sonnet/Haiku for lighter work) is both sufficient and more economical; routing only the genuinely hard steps to Fable 5 balances cost and quality. When we build a flow we match each step to the right model — not ‘use the most expensive one everywhere,’ but ‘get the job done at the right cost.’