For regular Assistant work across support and operations.
Included credits
225/ week
~975 average monthly credits
Billing mode
On-demand
Starts at
$20/mo
Features include
For teams running higher-volume Assistant workflows.
Included credits
2,250/ week
~9,750 average monthly credits
Billing mode
On-demand
Starts at
$150/mo
Features include
For heavy model use, larger contexts, and launch periods.
Included credits
9,000/ week
~39,000 average monthly credits
Billing mode
On-demand
Starts at
$500/mo
Features include
For contracted volume, governance, and managed rollout needs.
Included credits
Unlimited/ week
Usage based average monthly credits
Billing mode
Postpaid
Starts at
Custom
Features include
Applied handles the workspace, recovery, deployment, and day-to-day operating path so the assistant is ready for real work.
People, assistants, tools, reports, and automations follow the same permission model.
Every meaningful action can leave behind status, trace, source, and outcome context.
The assistant can investigate, report, draft, update, notify, and automate instead of only answering questions.
Assistant plans separate monthly subscription level from weekly credit budget, so teams can test, launch, and scale without exposing provider-specific units.
Assistant usage is deducted in whole credits per request, with usage rounded up after model work completes.
Included credits refresh every Monday at 00:00 UTC, so each plan has a predictable weekly operating budget.
On-demand plans use included credits first, then purchased top-up credits when usage exceeds the weekly allowance.
Enterprise plans do not hard-cap Assistant usage in-product. Usage is metered for contracted billing terms.
Rates are credits per 1M tokens. Cache reads are priced below full input where supported, and long-context premiums may apply to large premium requests.
| Model | Provider | Class | Context | Input | Output | Cache read |
|---|---|---|---|---|---|---|
| Credits per 1M tokens | ||||||
| gpt-5.5 | OpenAI | Premium | 1.05M | 1,000 | 6,000 | 100 |
| gpt-5.4 | OpenAI | Standard | 1.05M | 500 | 3,000 | 50 |
| gpt-5.4-mini | OpenAI | Economy | 400k | 150 | 900 | 15 |
| gpt-5-nano | OpenAI | Economy | 400k | 10 | 80 | 1 |
| claude-opus-4-7 | Anthropic | Premium | 1M | 1,000 | 5,000 | 100 |
| claude-sonnet-4-6 | Anthropic | Standard | 1M | 600 | 3,000 | 60 |
| gemini-3.5-flash | Gemini | Standard | 1.05M | 300 | 1,800 | 30 |
These estimates show approximate requests per 1,000 credits. Actual usage depends on model selection, context size, cache behavior, and the whole-credit minimum per request.
| Scenario | Request shape | Premium | Standard | Economy |
|---|---|---|---|---|
| Requests per 1,000 credits | ||||
| Short support reply | 1k input / 250 output | ~333 | ~500 | ~1,000 |
| Normal support turn | 2k input / 500 output | ~200 | ~333 | ~1,000 |
| Long support turn | 8k input / 1k output | ~71 | ~142 | ~333 |
Answers to credit and plan questions before you choose an Assistant plan.
Assistant credits are the unit used to price model usage. Each response records the credits deducted, and customer-facing rates are shown in credits instead of provider dollars.
Included credits refresh weekly on Monday at 00:00 UTC. Monthly credit numbers are averages for planning and comparison.
Launch, Growth, and Scale use purchased top-up credits after the included weekly credits are used.
Enterprise usage is postpaid and does not hard-cap in-product. Applied records usage for external contract billing.
Credit usage is rounded up per request, so very small requests are dominated by the minimum whole-credit charge.
Applied Assistant
Give it the tools, skills, and context your team uses every day. Start with free signup credits, then choose the plan that fits as usage grows.