Claude Opus 5 outperforms Fable 5 in coding at half the price
Anthropic's most aligned model to date hits the market.
1Launch of Claude Opus 5
Anthropic announced Claude Opus 5, which surpasses its previous flagship Fable 5 on most coding and knowledge-work benchmarks while halving the cost per token. As the fourth model released in just two months, Opus 5 offers developers a new option in the AI landscape.
2Performance Details
On Frontier-Bench, Opus 5 doubled its predecessor's score in building working software from engineering drawings. On ARC-AGI 3, it scored three times higher than the next-best model in novel problem-solving tests. These results demonstrate the model's superior capabilities.
3Business Applications
On Zapier AutomationBench, Opus 5 outperformed rivals in completing real business tasks end-to-end. Even at its lowest effort setting, the model successfully completed more tasks, positioning itself as an ideal tool for business process automation.
4Safety and Alignment
Anthropic states Opus 5 is its most aligned model to date, recording the lowest rate of deceptive behavior in automated safety audits. While setting new standards for honesty, it still trails Mythos 5 in offensive cybersecurity tasks.
5Pricing and Access
Opus 5 is priced at $5 per million input tokens and $25 per million output tokens. Available as the new default on Claude Max, the model can be accessed via API, Amazon Bedrock, Google Cloud, and Microsoft Foundry. A Fast mode option is also available.
6The Future of AI Market
Anthropic's rapid model releases indicate where the AI market is heading. The company now delivers four models targeting different price-performance segments in just eight weeks. Opus 5 emerges as a suitable option for everyday coding and knowledge work.
Original article: The Next Web — Claude Opus 5, Fable 5'i kodlamada yarı fiyatına geçti