What's new in Claude

Auto mode is now available on Bedrock, Vertex, and Foundry for Opus 4.7 and Opus 4.8 — opt in with CLAUDE_CODE_ENABLE_AUTO_MODE=1.

Plugins in .claude/skills directories now auto-load without the marketplace, and a new claude plugin init <name> command scaffolds new plugins with autocomplete for all subcommands.

Hotfix: resolved a thinking-block modification issue that occurred when using Opus 4.8 in extended thinking mode.

Claude Opus 4.8 is now the default model for Max, Team Premium, Enterprise pay-as-you-go, and Anthropic API accounts — high effort on by default, with /effort xhigh available for the hardest tasks.

Dynamic Workflows (research preview) let Claude write an orchestration script on the fly and spin up dozens to hundreds of coordinated subagents in parallel — results are verified before folding in, and the run iterates until answers converge.

Security-guidance plugin reviews Claude's own changes for vulnerabilities as it works — runs inline without interrupting your flow.

Fast mode on Opus 4.8 is now priced at $10/$50 per MTok and is the default for Max plan users, delivering up to 2.5× faster output generation.

Claude Managed Agents webhooks, multiagent orchestration, and self-hosted sandboxes are now available on Claude Platform on AWS, with new IAM actions and the AnthropicSelfHostedEnvironmentAccess managed policy.

Claude Opus 4.8 launched with a 1M token context window by default, 128k max output tokens, mid-conversation system messages, refusal categories in stop_details, and a lower 1,024-token minimum for prompt caching.

Adaptive thinking on Opus 4.8 triggers reasoning only when a turn needs it, cutting wasted thinking tokens vs. Opus 4.7 at the same effort level.

Workflows are available as a research preview in Claude Code, letting you define and run multi-step agentic plans. Fast mode for Opus 4.6 is now deprecated (removal ~30 days after launch).

Announced Dynamic Workflows in research preview — Claude plans dynamically, fans work out across subagents in parallel, agents refute each other's findings, and the run iterates until answers converge. Available on all paid Claude Code plans.

Claude Opus 4.8 is here — the most capable generally available Claude model, now the default on Max and Team Premium. Also: Claude Design doubled its monthly token limits on May 18, and the Code with Claude conference wrapped up in London.

"Hello from Code with Claude London!" — Boris Cherny (creator of Claude Code) posted from the London developer conference. He's also noted publicly that he hasn't written a single line of code himself in 2026.