AI Models in awork at a Glance
Configure which AI models your team can access by enabling or disabling Fast, Balanced, and Deep Thinking options in your workspace Models settings
awork AI supports various models from different providers. This means you're not locked into a single provider and can choose the model that fits your task, or use the models that awork has predefined for you.
As an admin, you control which models your team is allowed to use.
Benefits at a Glance
-
LLM Independence: awork supports models from various providers
-
Model Selection: Use different available AI models for different tasks
-
Admin Control: Workspace admins determine which models are available in the workspace
Available AI Models
awork supports various Large Language Models (LLMs) from different providers in different regions. You can choose between models from Anthropic, Google, OpenAI, and TensorX.
The models differ in terms of hosting and credit consumption, among other things. Not every model supports all features. Depending on the model, certain tools or agent functions may not be available, for example.
Important: awork doesn't offer every available model. One key factor is whether the model can be provided via a GDPR-compliant solution in the EU.
Managing AI Models in Your Workspace
Only Admins have access to AI model management and can disable or enable models available for use in the workspace.
-
Navigate to Settings (gear icon in the bottom left of the main menu)
-
In the AI category, select Models.
Configuring Predefined Models
Three predefined model categories are available: Fast, Balanced, and Deep Thinking.
- Fast for quick tasks
- Balanced for a balanced approach
- Deep Thinking for complex tasks
Each category has a dropdown menu for model selection.
By default, Auto is selected. In this case, awork decides the underlying model and appropriate reasoning level. The model used can change when models are updated or replaced. This keeps your selection automatically aligned with the recommended model. For everyday use, we recommend the Auto setting.
However, you can also choose a specific available model.
-
Click on the dropdown menu to the right of your desired category (Fast, Balanced, or Deep Thinking).
-
Choose between Auto and one of the available models.
-
Once you've selected a model, you can optionally adjust the reasoning level.
Restricting Allowed Models
In the Allowed Models section, you can specify which regions and providers should be available for your workspace.
Selecting Regions
Click on the Regions dropdown menu. Select your desired hosting regions from the list (Europe or US). Selected regions are marked with a purple checkmark. After selection, the field displays the number of selected regions (e.g., "1 selected").
Setting Providers
Click on the Providers dropdown field. The available LLM providers are:
- Anthropic
- OpenAI
- TensorX
Select your desired providers by clicking on the corresponding checkboxes. The field will then display the number of selected providers (e.g., "3 selected").
Once a region or provider is disabled, models from that provider cannot be individually enabled.
Automatically Enable New Models
Use the Automatically enable new models toggle to determine whether new models are automatically enabled for your workspace. This setting applies only to the regions and providers you've previously allowed.
Tip: If you want to consciously control available models in your workspace, you can disable automatic activation of new models. New models will then need to be manually enabled. This is particularly useful if you've intentionally excluded certain expensive models due to their higher credit consumption.
Managing Individual AI Models
In the AI Models section, you'll see a list of all available models with a counter showing how many models are enabled.
Enabling or Disabling a Model
Each model in the list displays the following information:
- Model name and provider (e.g., "GPT-5.6 Terra OpenAI")
- Hosting region (e.g., "Hosted in Europe")
- Approximate credit consumption (e.g., "~1.0x Credits")
The hosting region refers to the region from which a model is deployed, not the provider's country of origin.
Click the toggle to the right of your desired model to enable or disable it.
A disabled model cannot be used for AI operations in the workspace.
Selecting a Replacement Model When Disabling
If you disable a model that's still in use, a Select replacement models dialog will open. Here you must select a replacement model for each affected predefined model, agent, or thread before it can be disabled.
- Click on the Replacement dropdown menu next to the affected preset
- Select an alternative model from the list
- Confirm with Apply and disable or cancel the process with Cancel
Managing Image Models
Image models can be enabled and disabled separately from general AI models.
Disable models if you want to use only a specific image model.
If all image models are disabled, image generation in awork will no longer be available.
Credit Consumption Per Model
AI Credits are consumed based on the model you use.
This means: Not every model costs the same number of credits. Consumption is calculated relative to the base model GPT-5.6 Terra from OpenAI. The base model corresponds to a consumption of 1.0x Credits. Other models are ranked based on this value.
Some models consume significantly fewer credits, while more powerful or expensive models can consume a multiple of this amount.
Hover your mouse over the estimated consumption for more details.
For example, ~0.1x Credits means: For a comparable request, this model consumes 0.1 times as many credits as GPT-5.6 Terra from OpenAI.
Actual consumption may vary.
Currency of Available Models
awork generally keeps available models up to date and replaces outdated models with newer ones. An already selected and enabled outdated model can continue to be used, but is no longer available for new selection.
Note: There is no guarantee that a specific model or provider will remain available indefinitely.
Switching AI Models During a Thread
You can switch the predefined model (Fast, Balanced, and Deep Thinking) or selected model even during an ongoing thread.
Be aware: This may result in additional credit consumption because the prompt cache is reset when you switch models.
FAQ
Which AI providers does awork support?
awork is provider-independent and can use models from various providers, including OpenAI, Google, and Anthropic.
Does awork offer all available AI models?
No. awork decides which models and providers are supported. Available GDPR-compliant solutions and hosting regions are factors in this decision.
How current are the AI models in awork?
New models are generally integrated promptly. After a new model is released, it may take a few days before it becomes available in awork.
Who decides which models the team can use?
Workspace admins can define available models and disable hosting regions, providers, or individual models.
Can admins block specific models?
Yes. Admins can restrict individual models, providers, or regions.
Why do credit costs differ depending on the model?
The models have different consumption values. awork uses GPT-5.6 Terra with a factor of 1.0 as the base for credit calculation.
Can I switch models during an ongoing thread?
Yes. Please note: This may result in additional consumption because the prompt cache is reset when you switch models.
Are customer data used to train AI models?
No. Your data neither trains awork AI nor any model behind it.
Powered by happysupport.ai