Claude AI’s Opus 5 responses grow longer and more structurally complex compared to Fable 5

Anthropic released Claude Opus 5 on July 24, 2026, and users immediately noticed something: it talks a lot more than its predecessor. Responses from the new model are significantly longer, more detailed, and more structurally complex than what Opus 4.8 produced. Meanwhile, Fable 5, the company’s premium-tier model released earlier this year, keeps things tight and occasionally throws in a compliment for good measure.

What Opus 5 actually brings to the table

The new model ships with a 1 million token context window and a maximum output token capacity of 128,000 tokens. On the pricing front, Anthropic kept things simple. Opus 5 maintains the same API rate as Opus 4.8: $5 per million input tokens and $25 per million output tokens. That’s roughly half what Fable 5 costs, while delivering performance that closely rivals the top-tier model.

Opus 5 now serves as the default model on Claude Max and stands as the strongest model available on Claude Pro. It also comes with a Fast mode option for users who need quicker turnaround. The model’s tendency toward verbosity means users must explicitly prompt for brevity, a notable behavioral shift from Opus 4.8.

Fable 5: the quiet counterpart

Users have described Fable 5’s style as carrying a “quiet confidence,” delivering concise responses that often include praise or validation alongside the actual answer.

One tension to monitor: verbosity has costs, both literal and practical. Longer responses mean more output tokens consumed, which means higher bills for API users even at the same per-token price. Anthropic’s decision to require explicit prompting for conciseness puts the burden on users to manage that tradeoff.