Claude Opus 5 Launches, Competes with Fable 5 at Lower Cost

Anthropic released Claude Opus 5, positioning it as a cost-efficient alternative to Fable 5. The new model reportedly rivals Fable 5's capabilities at a lower cost. Opus 5 has been described as Anthropic's most powerful and safest AI model yet, although some have noted it's not a significant breakthrough in agentic coding performance.

Anthropic released Claude Opus 5, pitching it as a close match for rival model Fable 5's capabilities at roughly half the price . Reporting on the launch cited pricing near $5 per million input tokens and $25 per million output tokens for Opus 5, versus roughly $10 and $50 for Fable 5, a straightforward two-to-one cost gap that Anthropic is emphasizing over a claim of outright superiority.

Benchmark figures circulated around the launch show Opus 5 improving meaningfully over its own predecessor, Opus 4.8, on frontier reasoning tests, and landing within a narrow margin of Fable 5's best scores on select coding and computer-use benchmarks while using a fraction of the compute cost. Anthropic has positioned Opus 5 as the new default for everyday professional use, including coding, research, and office-style agent tasks, while continuing to point to Fable 5 as the stronger option for the longest, most autonomous multi-step workloads.

The launch underscores a broader shift in the frontier-model market, where labs are competing on price-per-token as much as raw capability. Enterprise buyers and developers evaluating AI tooling could weigh Opus 5's lower cost against workloads where task length and autonomy matter most, since the two models appear aimed at different use cases rather than a direct like-for-like replacement.