High-speed output mode for Claude Opus 4.6 that provides faster token generation without switching to a different model. Currently available as a research preview.
# Toggle fast mode from the prompt
/fast
# Or configure in settings.json
{
"fastMode": true
}- Uses the same Opus 4.6 model with accelerated output generation
- Billed as extra usage at a premium rate with separate rate limits
- Interactive toggle shows current status: "Fast mode ON" or "Fast mode OFF"
- When enabled, displays with the penguin icon (↯) and orange color theming
- Has its own rate limits; when reached, enters a cooldown period and automatically re-enables when reset
- Model automatically switches to Opus 4.6 if not already selected
Availability:
- Requires a paid subscription
- Requires extra usage billing to be enabled (
/extra-usage) - Requires the native binary (not available via npm install)
- Not available on Bedrock, Vertex, Foundry, or in the Agent SDK
- Organizations can disable fast mode for their accounts
Rate Limiting:
- Displays "Fast limit reached and temporarily disabled · resets in X" when hitting limits
- Shows "Fast limit reset · now using fast mode" when cooldown expires
- Tracks remaining time to reset in the UI
Promotions:
- May show discount percentages for extra usage (e.g., "50% off through [date]")
Fast mode system (search for "Fast mode", "tengu_penguins_enabled", "/fast")
Strings lifted out of the shipped bundle, so the claim above can be checked against them.
tengu_penguins_enabled On for this account, and not off by defaultThe flag server returned on for the one account this site reads, and nothing in this release compiles it off by default. The compiled default is shown below, and says which it is when we cannot read one: a fifth of gates compile in a string or a number rather than on or off, and most published releases have no gate table behind them at all. No client can see what the server returns for your account.
This account: on · anonymous baseline: on · compiled default in v2.1.36: no gate table built for this version
These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release.
Read once, for one account on one subscription tier, against v2.1.36. It isn't a statement about your account. What a flag value here can and cannot tell you
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.