Pricing
Not listed
Official monthly price not confirmed
- API
- Models are consumed via Azure AI Foundry and Azure OpenAI-compatible endpoints documented on Microsoft Learn
Finding the possibilities
Phi
● Identity checkedSmall language models · Vendor site 2026-10-04
Microsoft Phi is a family of small language and multimodal models designed for efficient on-device and cloud inference, offered through Azure AI Foundry alongside tools to build agents and applications. Phi models target developers who need capable models with lower compute footprints than frontier-scale systems.
- Category
- Small language models
- Free access
- Azure accounts include free-tier exploration options; Phi model usage bills through Azure subscription meters rather than a standalone Phi subscription
- API access
- Models are consumed via Azure AI Foundry and Azure OpenAI-compatible endpoints documented on Microsoft Learn
Is Phi right for you?
A good fit for
Developers and enterprises deploying compact models for copilots, edge scenarios, and cost-sensitive inference on Azure.
Before you choose
Pricing depends on chosen Azure deployment type and region; the Phi product page does not list flat monthly USD SaaS plans.
Overview
Phi brings Microsoft's small-model research to Azure customers who want lighter-weight language and vision models integrated with Foundry tooling instead of only frontier GPT-class endpoints.
What it can do
Features & capabilities
Unknown is different from unavailable. Each fact carries its own evidence.
| Capability | Value | Evidence | Checked |
|---|---|---|---|
| Model family | Azure markets Phi open models for text and multimodal scenarios within the broader Foundry catalog. | Facts sourced | 2026-10-04 |
| Deployment surface | Phi models appear among Foundry Models sold by Azure with region and deployment-type availability tables on Learn docs. | Facts sourced | 2026-10-04 |
Understand the total cost
Phi pricing & plans
Alternatives to Phi
View all ↗
CosineCosine, hosted at cosine.sh, is a sovereign AI lab training specialized coding agents and models for secure environments. It offers Lumen Scout, Outpost, and Sovereign models plus CLI and cloud surfaces with transparent credit-based plans for developers and regulated teams.
Explore tool
AppVeyorAppVeyor is a hosted CI/CD service for Windows, Linux, and macOS that connects to source control and runs builds on cloud VMs with admin access. It supports multi-stage deployments, self-hosted build agents, and parallel jobs for teams shipping cross-platform software.
Explore tool
Deno DeployPlans scale request quotas, egress, custom domains, and collaboration features for hobby projects through enterprise deployments.
Explore toolThe practical questions
Frequently asked questions
What is Phi used for?
Microsoft Phi is a family of small language and multimodal models designed for efficient on-device and cloud inference, offered through Azure AI Foundry alongside tools to build agents and applications. Phi models target developers who need capable models with lower compute footprints than frontier-scale systems.
Does Phi have a free plan?
Azure accounts include free-tier exploration options; Phi model usage bills through Azure subscription meters rather than a standalone Phi subscription. This record does not confirm an ongoing free plan.
How much does Phi cost?
A listed monthly price has not been confirmed. See the plan cards for entitlements, billing commitments and seat minimums.
Can I use Phi through an API?
Models are consumed via Azure AI Foundry and Azure OpenAI-compatible endpoints documented on Microsoft Learn. API access and subscription access may have different terms; consult the linked sources.
What should I check before choosing it?
Pricing depends on chosen Azure deployment type and region; the Phi product page does not list flat monthly USD SaaS plans.
Price history
No retained pricing changes yet. A current price alone does not establish a historical trend.
How this profile is supported
Identity and profile
Reviewed vendor source
Read original source ↗Facts apply to the named version and check date. Send a sourced correction if something changed.