Meet the New Claude Opus 5: Frontier-Class Agentic Coding and Computer Use at Unchanged Opus Pricing

TL;DR AI
2 min readKey summary
Anthropic has released Claude Opus 5 as its new flagship Opus-tier model, with pricing unchanged at $5 per million input tokens and $25 per million output tokens.
The model now defaults to thinking on, uses an effort setting to control reasoning depth, and introduces new request rules when thinking is disabled.
Claude Opus 5 ships with a 1M-token context window and higher output limits, making it more capable for long, complex workloads.
Anthropic says the model delivers major gains in coding, computer use, multimodal tasks, and reasoning, with stronger results on SWE-bench, OSWorld, and ARC-AGI-3.
The upgrade boosts performance without raising price, but developers will need to adapt to API-breaking behavior changes.
