Claude Opus 4.8 Launches With Fast Mode, Dynamic Workflows, and New Effort Controls.

Anthropic released Claude Opus 4.8 today, upgrading its flagship model with improved reasoning, longer autonomous work sessions, and the ability to better assess its own progress. The new version ships at the same price as its predecessor, Opus 4.7, and is available immediately across claude.ai, the Claude Platform, and major cloud providers.

The launch introduces three significant additions: Fast mode for Opus 4.8 in research preview, dynamic workflows with parallel subagents in Claude Code, and an effort control that lets users adjust how much computational reasoning Claude applies to each task.

Fast mode runs the same Opus 4.8 model at roughly 2.5 times the speed of the standard version, with pricing now three times cheaper than before. Anthropic positioned the mode as suitable for tasks where latency matters more than maximum depth, following their recent partnership with SpaceX to expand compute capacity.

Dynamic workflows, also in research preview, allow Claude Code to spin up hundreds of parallel subagents within a single session. The system verifies its own work before reporting results, a shift that lets developers hand off larger scopes like feature implementations, database migrations, or bug sweeps without manual checkpoints.

The parallel architecture differs from sequential task execution in tools like Cursor or Windsurf, where agents typically operate in series. Anthropic's approach distributes subtasks across multiple instances that run simultaneously, then synthesizes results before presenting them to the user.

The new effort control on claude.ai lets users dial response intensity up or down. Lower settings produce faster, lighter answers for straightforward queries. Higher settings allocate more reasoning steps for complex problems that benefit from extended thought.

This parallels adjustments other platforms have introduced for balancing speed and depth. ChatGPT offers similar toggles through its model selector, while Gemini exposes temperature and token controls in its API.

The control does not appear to affect pricing directly, but does influence token consumption and response time, which indirectly impacts cost for users on usage-based plans.

Anthropic described Opus 4.8 as having "sharper judgment" and "more honesty about its own progress," language that suggests improvements in metacognitive capabilities rather than raw benchmark performance. The model can now work independently for longer periods, reducing the frequency of clarification requests during complex tasks.

The company did not publish new benchmark scores alongside the release, focusing instead on practical workflow improvements in Claude Code. Developers using the platform can now assign multi-step projects and return later to reviewed, validated output.

Claud Opus 4.8 is live today across all supported platforms. Fast mode and dynamic workflows remain in research preview, with broader availability expected pending feedback from early users. Pricing for the standard model remains unchanged from Opus 4.7.