# Phi-4-multimodal API prices

Phi-4-multimodal (phi) — 128k context, 16k max output.
3 metered per-token offers from 3 provider(s), in USD per 1M tokens.
Prices as of 2026-10-04.

| Provider | Input $/1M | Output $/1M | Cache read $/1M | Context |
| --- | --- | --- | --- | --- |
| NanoGPT | $0.07 | $0.11 | $0.035 | 128k |
| Azure | $0.08 | $0.32 | — | 128k |
| Azure Cognitive Services | $0.08 | $0.32 | — | 128k |

## Summary

- Cheapest input: $0.07 per 1M tokens (NanoGPT)
- Cheapest output: $0.11 per 1M tokens (NanoGPT)
- First-party: not listed separately
- Free offers: none
- Subscription plans (not per-token): none
- Released: 2024-12-11
- Inputs: audio, image, text

## FAQ

### What is the cheapest Phi-4-multimodal API?

As of Oct 4, 2026, NanoGPT has the lowest Phi-4-multimodal output price at $0.11 per 1M tokens, and NanoGPT has the lowest input price at $0.07 per 1M tokens. Prices are metered per-token rates in USD; free tiers and subscription plans are excluded.

### How many providers offer Phi-4-multimodal?

3 providers list Phi-4-multimodal on Sovyron; 3 of them sell it at a metered per-token price.

### Is Phi-4-multimodal free?

No provider in the Sovyron catalog lists a free tier for Phi-4-multimodal.

### What is the context window of Phi-4-multimodal?

Phi-4-multimodal supports a 128k-token context window and up to 16k output tokens.

HTML page: https://sovyron.com/models/phi-4-multimodal/
Full dataset: https://sovyron.com/data/catalog.json