Update
xAI Grok Models Are Now Available in Dawn
Dawn now supports xAI as a model provider, making Grok models selectable for agents alongside the rest of the Dawn model catalog.
Dawn Team
Product Engineering
Dawn now supports xAI as a model provider, bringing Grok models into the Dawn model catalog. Workspace agents can use Grok alongside the other model families already available in Dawn.
Provider choice in Dawn is per agent. That means a workspace can use different model families for different jobs: a fast model for routine chat, a deeper reasoning model for review, a multimodal model for file-heavy work, and Grok where the team wants to evaluate xAI’s models on real workspace tasks.
Choose Grok for the Agent That Needs It
Model choice should follow the work. Dawn does not force the whole workspace onto one provider, and this release keeps that pattern.
If you want to test Grok on code review, create or edit a review agent and select the Grok model from the model picker. If you want to compare it on planning, make a planning agent and keep the rest of the workspace unchanged. If you want one experimental agent for evaluation, create that agent without touching production workflows.
That makes evaluation practical. Teams can try Grok on real work without disrupting every other agent in the workspace.
Same Dawn Features Around the Model
Selecting a Grok model does not change the way the rest of Dawn works. The agent still uses its prompt, assigned skills, memory access, connected sources, files, and channel bindings. The model changes; the workspace workflow stays familiar.
That is the value of adding providers to the catalog rather than treating each provider as a separate product surface. A team can compare model behavior while keeping the same agent setup and the same day-to-day Dawn experience.
Useful for Provider Diversity
Some teams want provider diversity because different models have different strengths. Others want it for cost control, availability, policy, or procurement reasons. A workspace might use Dawn-managed credentials for common work, bring its own key for a specific provider, or keep a few agents on different model families to compare behavior.
Grok support gives those teams another option. It also makes Dawn a better place to evaluate models in context, because the test happens against real prompts, real sources, and real workflows rather than isolated benchmarks.
What to Expect
As with any model change, behavior can differ. A Grok-backed agent may write differently, choose tools differently, or need prompt adjustments compared with an agent using another provider. That is normal and useful to evaluate.
Start with a focused agent and a repeatable job. Ask it to review a pull request, summarize a source, prepare a plan, or answer a question with connected context. Compare the result with the model you already use for that job, then decide whether the fit is right.
Getting Started
Open Dawn Web, edit the agent you want to try, and choose an xAI Grok model in the model picker. If your workspace uses provider keys, configure xAI in the same provider-management area where you manage other model providers.
The safest rollout path is to add one Grok-backed agent first. Use it on real work, compare the result, and only move other agents when the model is earning its place.