Executive Summary
AI Gateway has launched "Fast mode," an experimental premium feature for the Claude Opus 4.6 model. This mode increases output token generation speed by 2.5x without compromising the model's intelligence, targeting developers with latency-sensitive tasks like complex coding and planning. The feature is available immediately but is priced at 6x the standard rate for both input and output tokens.
Key Takeaways
* Product: Fast mode for Claude Opus 4.6.
* Primary Function: A high-speed option that provides 2.5x faster output token speeds.
* Target Use Cases: Human-in-the-loop workflows, large-scale coding tasks, and planning that benefit from reduced wait times.
* Status: The feature is considered early and experimental.
* Implementation: Enabled by passing a `speed: 'fast'` parameter in API calls or setting `"fastMode": true` in configuration.
* Pricing: Costs 6x standard rates, at $30 per million input tokens and $150 per million output tokens.
* Availability: Available now through the AI Gateway, including a direct playground integration.
Strategic Importance
This move introduces a premium performance tier, allowing the company to monetize low-latency needs and unlock new, more interactive use cases for its most powerful models.