Anthropic Claude Opus 4.8 launches Ultracode Mode for deeper reasoning

Anthropic has released Claude Opus 4.8, the latest version of its flagship AI model, with a focus on coding, reasoning, and transparency. The headline feature is user-controlled 'effort' settings, informally called Ultracode Mode by developers. This allows users to dial the model's computational intensity from fast, low-cost responses to deep, prolonged reasoning for complex coding and debugging tasks. Higher effort uses more tokens but improves performance on software engineering, planning, and analytical work. Additionally, Anthropic introduced Dynamic Workflows, a research-preview feature that enables Claude to coordinate hundreds of AI sub-agents simultaneously—dividing tasks, verifying outputs, and synthesizing results. The model also shows improved honesty: it is four times less likely than its predecessor to ignore coding errors without warning, reducing the risk of delivering incorrect or unsupported answers.

Compared to OpenAI's GPT-5.5, Claude Opus 4.8 is positioned as a strong contender in coding and long-duration agent tasks. Early tester feedback and benchmarks suggest it may outperform GPT-5.5 in selected programming scenarios, though GPT-5.5 retains advantages in multimodal capabilities and a broader ecosystem. Pricing remains unchanged despite the upgrades, and Anthropic added cost-control options for simpler tasks. Safety is also highlighted: while jailbreak resistance has improved, the company acknowledges vulnerabilities to sophisticated attacks. The launch signals a shift toward more customizable, honest AI assistants tailored for professional developers and enterprises.