Claude Opus 4.8 is here: effort controls, dynamic workflows, cheaper fast mode, better honesty, less deception

Anthropic launched Claude Opus 4.8, the newest flagship Claude model, with new effort controls and a cheaper fast mode.

The update adds research-preview dynamic workflows for large coding jobs in Claude Code and aims to improve honesty while reducing deception.

Anthropic says Opus 4.8 beats Opus 4.7 on benchmarks and outperforms GPT-5.5 and Gemini 3.1 Pro on several measures.

The company still trails OpenAI on agentic terminal coding, but the release strengthens Claude’s competitiveness in coding and safety.