Anthropic announces Claude Opus 4.8, an upgrade from Opus 4.7 with improved coding performance and honesty

Anthropic has launched Claude Opus 4.8, an upgraded model over Opus 4.7 with better agentic coding, reasoning, computer use, and financial analysis.

The model reportedly delivers stronger benchmark results and is positioned as more honest and reliable, with safety improvements alongside performance gains.

Anthropic also introduced new Claude Code dynamic workflows, effort controls in Claude apps, and a Messages API update for changing agent context.

The release strengthens Anthropic’s top-tier Claude offering for developers building long-running AI agents and coding workflows.