Claude Opus 5: Anthropic's Best & Most Cost-Effective AI

Anthropic has launched Claude Opus 5, calling it their best performing and most cost-effective AI model yet. It outperforms its predecessor, Opus 4.8, in coding and knowledge work, and is more capable than Fable 5, Anthropic's most advanced publicly available model, at about half the cost. Opus 5 is priced at $5 per million input tokens and $25 per million output tokens. This matches the previous cost of Opus 4.8 while offering a significant performance boost. The model shows strong results in software engineering and business automation. On one coding benchmark, it more than doubled its predecessor's performance for less cost per task. It also scored roughly three times higher than the next best model in novel problem-solving. Opus 5 also shows gains in scientific research, including improvements in organic chemistry and protein analysis. However, Opus 5 is not designed for high-risk capabilities like cybersecurity. It lags behind Mythos 5, Anthropic’s most restricted system, in identifying software vulnerabilities. This gap is intentional, as Opus 5 was not trained on offensive cyber tasks. Opus 5 also trails Mythos 5 in biology-related research, an area where Anthropic applies extra caution. This launch reflects a broader market shift, with businesses seeking clear return on investment from AI models. This matters because it indicates a growing focus on value and efficiency in the rapidly evolving AI industry.