Urgent.News

What's breaking now, across thousands of outlets.

AI

Claude Opus 5.5 Effort Levels Tested: Low to Max

I'm a solo developer in Korea who builds games with AI. Claude Opus 5.5 has five effort levels, so I ran the same prompt at every one of them and measured it. Claude Opus 5.5 has an effort setting that controls how deeply it thinks, with five levels: low, medium, high, xhigh and max. Higher levels are supposed to give better results, but the official docs give no numbers for how much extra time…

Claude Opus 5.5, an AI model, offers five levels of effort, ranging from low to max, each with varying costs and processing times. The model's default setting is medium effort, but users can manually adjust it using the --effort option in Claude Code or by setting the effort value in the API. To test the model's performance at each effort level, a game development prompt was run five times, with results measured in terms of time taken, output tokens, total tokens, and cost.

The low effort level took only 32 seconds and cost $0.48, generating a simple brick-breaker game with basic features. Medium effort, the default setting, cost $0.56 and took 52 seconds, adding rainbow bricks, a pause button, and level-clear screens. High effort cost $0.75 and took 1 minute and 50 seconds, introducing particle effects, a saved best score, and high-quality design.

Xhigh effort, costing $1.36 and taking 4 minutes and 32 seconds, added sound effects, level names, screen shaking, and victory fireworks. Max effort, the most expensive at $5.25 and taking 22 minutes and 22 seconds, included sound effects, level names, an invader-shaped level two, screen shake, and victory fireworks.

Despite the increased time and cost, higher effort levels resulted in more polished and feature-rich games. The differences in output quality became more apparent as the effort level increased, with max effort providing the most advanced game-building capabilities. Low effort was suitable for quick prototypes and testing, while high and max effort were better suited for polished, feature-rich games.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

RAG vs AI Agents: Understanding the Difference Through Practical Examples

Large language models (LLMs) can answer questions, summarize documents, and generate code. However, building a reliable AI application often requires more than sending a prompt to a model.

  • RAG focuses on locating and integrating external information for accurate answers
  • AI agents coordinate tools and multi-step workflows to accomplish goals
  • Hybrid RAG and AI agent approach combines strengths for comprehensive solutions

More from Friday 9 October →