Overview
- Opus 4.5 is available now in Claude apps, via API, and on AWS, Google Cloud and Azure, and it becomes the default for Pro, Max and Enterprise plans.
- Anthropic reports an 80.9% SWE‑Bench Verified score, edging OpenAI’s GPT‑5.1‑Codex‑Max (77.9%) and Google’s Gemini 3 Pro (76.2%).
- Opus‑tier API pricing drops to $5 per million input tokens and $25 per million output tokens, down from $15 and $75 respectively.
- Efficiency and UX updates include an effort parameter for cost–performance control and longer‑conversation context management to reduce token use and hard stops.
- Product integrations expand with Claude for Chrome access, Claude for Excel general availability, and Claude Code on the desktop, as safety evaluations show 100% refusals in one agentic coding test but lower rates for malicious coding (~78%) and computer‑use (~88%).