Anthropic Releases Claude Opus 5 with Near-Flagship Performance at Half the Price
Anthropic has launched Claude Opus 5, a new flagship AI model that delivers performance close to the higher-tier Claude Fable 5 at half the cost, while keeping the price identical to the previous Opus 4.8. The model sets new performance records on benchmarks like ARC-AGI-3, Frontier-Bench v0.1, and CursorBench 3.2. This release significantly lowers the cost barrier for frontier-class AI capabilities, enabling developers to deploy highly advanced reasoning and coding agents more economically. It also demonstrates substantial progress in AI fluid intelligence and autonomous problem-solving. Claude Opus 5 is priced at $5 per million input tokens and $25 per million output tokens, with a "Fast" mode option that runs 2.5 times faster at double the price. Additionally, it triggers safety classifiers 85% less often than Fable 5, though it still blocks high-risk tasks like binary vulnerability exploitation.
## BACKGROUND
The Abstraction and Reasoning Corpus (ARC-AGI) is a benchmark designed to measure an AI's fluid intelligence and abstract reasoning through visual grid puzzles. Meanwhile, benchmarks like CursorBench and Frontier-Bench evaluate the capabilities of AI agents in executing complex, multi-file coding and software engineering tasks.