Choosing a Group (read this before creating a token)

When you create an API Key (token) in the console, you'll be asked to pick a group. The group determines, for that token:

• Which models you can use
• What price you're charged

Most of the time: just keep the default "auto" group

New tokens default to the auto group: each request is matched to the right route for the model you ask for, so one token covers GPT / Codex, Claude, Gemini and Grok — no need to choose.

These need their own group

• 🎨 Image generation → you must create a new token under the media / media-gen group
• GLM, Kimi, Muse and other models outside the auto group → pick the matching group when creating the token

Using the wrong group results in a "no available channel" error.

How to choose a group

1. Open Console → Tokens and click Add Token
2. The Group field defaults to the auto group; for the model types listed above, switch the dropdown to the matching group
3. After saving, the token matches models and is billed according to the group you chose

Want to pin a token to one specific route (say, a particular Claude group)? Just pick that group in the dropdown. If you have several different uses, creating several tokens, each with its own group, keeps things clearest.

Live demo: when creating a token, the 'Group' dropdown shows all groups (each with its own rate multiplier)

Which group is for what

What you want to use	Pick this group
GPT / Codex, Claude, Gemini, Grok	Auto group (default); or pick that vendor's own group
Zhipu GLM	GLM group
Moonshot Kimi	Kimi group
Meta Muse	Muse group
🎨 Image generation	media / media-gen

The console dropdown is authoritative

The exact group names and price multipliers are whatever the "Group" dropdown on the console token page actually shows (we add or adjust them from time to time). One vendor may have several groups with different price and stability tiers. To see whether each group is healthy right now, check the status page.

How the auto group is billed

The auto group has no price of its own — you are billed at the rate of whichever group the request actually lands on. The console usage log shows which group each request used.

Common error: no available channel

A "no available channel" / "no available channel under the current group" error means the token's group doesn't contain the model you requested.

How to fix:

• Confirm the model name is spelled correctly (see Quickstart)
• The token uses the auto group but still gets this error → the model is outside the auto group's range (e.g. image generation, GLM, Kimi); create a new token under the matching group
• Image generation → must be the media / media-gen group

For more, see Common Errors & Fixes.