Anthropic has officially released Claude Opus 5, delivering frontier-level AI performance and state-of-the-art agentic capabilities at half the cost of top-tier models.
Designed as a proactive daily workhorse, Opus 5 brings massive upgrades to software engineering, complex reasoning, and scientific research. On key benchmarks like Frontier-Bench, ARC-AGI 3, and OSWorld 2.0, it significantly outperforms its predecessor (Opus 4.8) and approaches the capability of Claude Fable 5. The model excels at long-horizon coding tasks, self-verifying bug fixes, legal analysis, and automated enterprise workflows.
Opus 5 maintains an accessible price point of $5 per million input tokens and $25 per million output tokens, alongside a Fast mode operating at 2.5× speed. Furthermore, Anthropic reports that Opus 5 is its most aligned and safe model yet, achieving lower rates of deceptive behavior and improved safeguard precision.
Claude Opus 5 is available today across Claude.ai, Claude Code, and the Claude API (claude-opus-5).
Source: Anthropic news