Anthropic releases Claude Opus 5

Sources: Anthropic announcement, Opus 5 system card, HN 49038433
Summary: Anthropic released Claude Opus 5 on 2026-07-24 as its new flagship model, API id claude-opus-5, priced at $5 per million input tokens and $25 per million output tokens, the same as Opus 4.8. It is available in the Claude API, on Claude.ai, and in Claude Code, with an optional fast mode at 2.5x speed for 2x the base price. Anthropic reports state-of-the-art results on its Frontier-Bench and GDPval-AA coding evaluations, more than double the Opus 4.8 score on Frontier-Bench v0.1, about three times the next-best model on ARC-AGI 3, and stronger OSWorld 2.0 computer-use results than Fable 5 at lower cost. The system card states Opus 5 is Anthropic's most aligned model to date by automated behavioral audit while still ranking behind Mythos 5 on cybersecurity exploitation and biology, with safeguards similar to Opus 4.8. All benchmark figures are vendor-reported and not independently reproduced.
Why it matters: Opus 5 becomes Anthropic's default frontier model for coding and agentic work at unchanged Opus pricing, resetting the price-performance baseline developers use for Claude Code and agent harnesses.
Follow-up: Watch for independent benchmark reproduction, whether Claude Code moves its default to Opus 5, and comparisons against Fable 5, GPT-5.6 Sol, and Kimi K3 on real coding tasks.