Maximize Logic on Massive Datasets with Opus 5 Effort Levels
Use 'max' effort for your most complex, 1M token reasoning tasks.
Set your Claude API or agent configuration to 'max' effort when performing deep reasoning or agentic work involving up to 1 million tokens.
The Scenario
You need to analyze dozens of long legal contracts or entire codebases simultaneously to find inconsistencies. You want the highest possible accuracy without splitting files.
Before & after
Previously, you had to manually chunk documents or use complex RAG systems to fit narrow windows. This could take 1–2 hours to verify all information was captured correctly.
Using Claude Opus 5 with the 'max' effort setting handles these 1M token tasks natively with deep reasoning. This reduces your time to 5–10 minutes of prompt setup and monitoring.
The Prompt
I am providing a large set of documents below. Please analyze them for [SPECIFIC_INSIGHTS] and provide a detailed report. Use the highest reasoning effort to ensure no details are missed across the entire 1M token context. [PASTE_DOCUMENTS_OR_TRANSCRIPTS_HERE]
Claude Opus 5 offers a step-change in reasoning and supports a massive 1-million-token context window by default, making it ideal for deep reasoning and long-horizon tasks.
Source
Claude Platform release notes - Claude Platform Docs"Claude Opus 5 supports a 1M token context window... and thinking on by default... with max for capability-critical work."