Claude Opus 5 Arrives: Frontier AI at Half Price?

Opus 5: A Leap in AI Capabilities

Anthropic's announcement of Claude Opus 5 marks a significant advancement in the large language model (LLM) landscape. The key takeaway is its impressive performance gains, particularly in coding and knowledge work, where it claims 'new state-of-the-art' status on several evaluations. The ARC-AGI-3 score, being three times higher than the next best model, is a particularly striking claim, suggesting a remarkable leap in novel problem-solving capabilities. Furthermore, the emphasis on improved alignment, with lower rates of reckless or deceptive behavior and stronger adherence to Claude's Constitution, addresses a critical concern in AI safety and responsible development. The pricing strategy, matching Opus 4.8 while delivering superior performance, is a bold move that directly challenges competitors and democratizes access to advanced AI for a wider range of users on paid plans and API access. The introduction of a 'Fast mode' further enhances usability and efficiency, catering to different application needs.

However, several nuances warrant closer examination. While Opus 5 is positioned as a competitor to 'frontier intelligence,' the comparison to 'Fable 5' is vague and lacks specific benchmark details, making it difficult to independently verify the 'half the price' claim relative to true frontier models. The cybersecurity disclosure, while transparent, highlights that Opus 5, despite being stronger than Opus 4.8, still 'substantially behind Mythos 5 at developing exploits.' This suggests a deliberate design choice to balance capability with safety, which is commendable but also implies limitations for certain advanced security research use cases. Developers seeking to push the boundaries of exploit development might find this a constraint. The 'automated behavioral audit' is a positive step, but the details of this audit and how 'reckless or deceptive behavior' are quantified are not provided, leaving room for deeper scrutiny. The immediate user demand for usage resets, as seen in the replies, indicates a strong desire to test the new capabilities, underscoring the developer appeal.

Key Points

• Claude Opus 5 is now available on all paid plans and the Claude API.
• It achieves new state-of-the-art performance on several coding and knowledge work evaluations.
• Opus 5 demonstrates significantly improved novel problem-solving capabilities, as evidenced by its ARC-AGI-3 score.
• The model is highlighted as being more aligned with lower rates of reckless or deceptive behavior.
• It is priced the same as Opus 4.8, offering a significant performance upgrade at no additional cost.
• A 'Fast mode' is available, offering approximately 2.5x the default speed.
• While stronger in cybersecurity tasks than Opus 4.8, it remains behind 'Mythos 5' in exploit development, indicating a focus on safety safeguards.

---

📖 Source: Opus 5 is available today on all paid plans and the Claude API, priced the same as Opus 4.8. It’s th...