Anthropic ends fast mode for Claude Opus 4.8 on November 30

Anthropic deprecated fast mode for Claude Opus 4.8 on October 30, 2026, and will remove it on November 30. After that date, a request to claude-opus-4-8 with speed: "fast" returns an error. Anthropic’s release notes tell users who want to keep fast mode to move to Claude Opus 5.5, where it costs $8 per million input tokens and $40 per million output tokens.

Fast mode runs the same model on a faster inference setup. Anthropic says it delivers up to 2.5x more output tokens per second, with no change to the model’s intelligence. The gain is in output speed, not time to first token.

What changed

On Opus 4.8, fast mode was a research preview on the Claude API that needed the fast-mode-2026-02-01 beta header. That ends on November 30. Standard-speed requests to Opus 4.8 are not part of this notice: the release note covers fast mode only, and it gives no retirement date for the model itself.

What it costs

Fast mode is a premium on top of standard rates, applied across the full context window. Prompt caching and data residency multipliers apply on top of it.

Fast mode, per 1M tokens	Input	Output	Status
Claude Opus 4.8 (until Nov 30)	$10	$50	Research preview
Claude Opus 5	$10	$50	Research preview
Claude Opus 5.5	$8	$40	Available

So moving fast traffic from Opus 4.8 to Opus 5.5 lowers the fast mode rate by 20% on both input and output. Opus 5.5 fast mode needs no beta header and works on the Claude API, Claude Models on Google Cloud (Anthropic managed) and Microsoft Foundry. Anthropic says Opus 5.5 has a default fast mode rate limit in the Start, Build and Scale usage tiers, so there is no access request. Fast mode on Opus 5 still needs a waitlist or an account manager.

What to do

Search your code for claude-opus-4-8 paired with speed: "fast" and plan the switch before November 30. Anthropic links a migration guide from Opus 4.8 to Opus 5.5. Keep in mind that fast mode is not available with the Batch API or a Priority Tier commitment, and that falling back from fast to standard speed causes a prompt cache miss.

To compare per-request costs before you switch, try our AI token cost calculator, or see current list prices on the AI models page.