Claude Opus 4.8 Workflows: Lower Cost, Cap at 1k Subagents

Anthropic released Claude Opus 4.8 along with two Claude Code updates: dynamic workflows and a cheaper fast mode. Dynamic workflows let Claude write a JavaScript script that orchestrates dozens of subagents running in parallel. The script lives outside Claude’s context window so only the final answer returns to your chat, keeping the session responsive while agents work. You can start a workflow by typing the word workflow anywhere in your prompt, enabling ultracode, or running the built‑in deep‑research command. The runtime limits concurrency to 16 agents and caps the total at 1 000 per run, and the script cannot touch the filesystem or shell—only the agents do. Progress is saved, so an interrupted job resumes from where it left off.

Fast mode is not a different model; it is a high‑speed configuration of Claude Opus that delivers 2.5× faster output token speed while keeping the same intelligence. You toggle it with /fast and a small icon marks the active session. For Opus 4.8 fast mode is three times cheaper than before, priced from usage credits rather than your plan’s included quota. It is best for rapid iteration, live debugging, or any situation where latency matters more than cost.

Both features consume more tokens than a typical chat, so start with a scoped task, monitor usage, and verify outputs before trusting long runs. Workflows excel at codebase‑wide migrations, bug hunts, and cross‑checked research where a single pass cannot reach sufficient confidence. Fast mode suits interactive work, quick prototyping, and iterative testing when you need speed without sacrificing model quality.

#AI #Product #Claude #LLM #DeveloperTools #AIEngineering