$0.056 / $0.11per 1M tokens in / out
#310on OpenRouter by tokens · 1.3B / week
Price sheetper 1M tokens
| Input | $0.056 | DeepInfra |
| Output | $0.11 | |
| Context · max output | 16K · 15K | tokens |
Where to buy3 providers · per 1M tokens
| Provider | Input | Output | Context |
|---|---|---|---|
| DeepInfra | $0.056 | $0.11 | — |
| Featherless | $0.07 | $0.14 | — |
| Azure AI Foundry | $0.13 | $0.50 | — |
In Microsoft's line-upper 1M in / out
| Microsoft Decision 1 | $0.042 / — | −100% |
| Phi 4 | $0.056 / $0.11 | this model |
| Phi-4 Mini | $0.075 / $0.30 | +87% |
| Phi 4 Reasoning Plus | $0.07 / $0.35 | 2× |
| Phi 3 Medium 4K Instruct | $0.14 / $0.14 | 2× |
| WizardLM 2 8x22B | $0.48 / $0.48 | 6.9× |
| Phi 3 Mini 4K Instruct | $0.40 / $0.80 | 7.1× |
Cheaper for the same workvs Phi 4
| Qwen3.5 4B thinkingOpen | -14% | score ×2.2 | $0.03 / $0.15 |
| Nemotron 3.5 Lightning thinkingOpen | -30% | score ×2.2 | $0.079 |
| GPT OSS 120B highOpen | -14% | score +96% | $0.092 |
| GPT OSS 20B lowOpen | -49% | score +68% | $0.018 / $0.09 |
| Granite 4.2 3B thinkingOpen | -25% | score +53% | $0.006 |
Pricing questions
How much does Phi 4 cost?›
$0.056 per 1M input tokens and $0.11 per 1M output tokens.
Is Phi 4 free to use?›
There is no free API tier, but the weights are open: you can run it on your own hardware for the cost of compute.
What is a cheaper alternative to Phi 4?›
Qwen3.5 4B at thinking effort costs -14% per task with a score ×2.2 on Humanity's Last Exam.