Context
16K
Max output
16K
Serving providers
3
Cheapest input
$0.07 /1M
On-demand listed rates, OpenRouter routed quotes and Vercel reseller quotes. Enterprise commitments (committed-use discounts, provisioned throughput, negotiated rates) can be materially lower.
Benchmarks
independent evaluations · source: Artificial AnalysisSpeed & latency
Output speed
42 t/s
tokens / second
Time to first token
2.46s
latency
Headline indices
Individual benchmarks · click a row for effort variants
| Benchmark | Best score | Effort | Cost / task | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CritPt | 0% | default | — | |||||||||||||
| EvalComputeProxyFrozen Aug 18, 2026 | 5.0 | default | — | |||||||||||||
| GPQA Diamond | 57% | default | — | |||||||||||||
| ||||||||||||||||
| Humanity's Last Exam | 4% | default | — | |||||||||||||
| IFBench | 24% | default | — | |||||||||||||
| LiveCodeBench | 23% | default | — | |||||||||||||
| Long-Context Reasoning | 0% | default | — | |||||||||||||
| MMLU-ProFrozen Aug 18, 2026 | 71% | default | — | |||||||||||||
| Omniscience | -55.7 | default | — | |||||||||||||
| Omniscience Accuracy | 0.1 | default | — | |||||||||||||
| Omniscience Non Hallucination | 0.2 | default | — | |||||||||||||
| SciCodeArtificial Analysis did not report this score in its latest run on Oct 14, 2026 | 26% | — | — | |||||||||||||
| TerminalBench Hard | 4% | default | — | |||||||||||||
| τ²-bench | 0% | default | — | |||||||||||||
Artificial Analysis last published scores for this benchmark on Aug 18, 2026. Scores here are frozen at that date, and models released since are not on it.
Pricing by serving provider
standard tier · per 1M tokens · click a provider to expandInput pricing runs from $0.07 to $0.125 per 1M tokens across 3 providers, so the dearest route costs 79% more than the cheapest for the same model.
| Serving provider | Input /1M | Output /1M | Endpoints | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DeepInfraCheapest | $0.07 | $0.14 | 1 | ||||||||||
Standarddeepinfra/microsoft/phi-4
| |||||||||||||
| OpenRouter | $0.07 | $0.14 | 1 | ||||||||||
Standardmicrosoft/phi-4
| |||||||||||||
| Azure | $0.125 | $0.50 | 1 | ||||||||||
Standardazure_ai/Phi-4
| |||||||||||||
Price history
listed seller prices, input and output per 1M tokens Input Output
Input down 44% since first tracked
Cost calculator
estimate your monthly spend on this model$25
estimated / month
Model IDs
copy the exact identifier for your platformdeepinfra/microsoft/phi-4microsoft/phi-4azure_ai/Phi-4
Frequently asked questions
Phi 4 pricing, context and availabilityHow much does Phi 4 cost?
Phi 4 costs $0.07 per 1M input tokens from DeepInfra, and $0.14 per 1M output tokens from DeepInfra. Across 3 serving providers, input prices range from $0.07 to $0.125 per 1M tokens.
What is the context window of Phi 4?
Phi 4 accepts up to 16K tokens of context and can return up to 16K output tokens.
Which providers serve Phi 4?
Phi 4 is available from 3 serving providers, each with its own pricing and model ID. DeepInfra is currently the cheapest.
What can Phi 4 do?
Phi 4 supports tool use.
Other Microsoft models
compare pricing across the Microsoft lineupWizardLM-2 8x22B$0.48/1M · 3 providersPhi 3 Mini 128k Instruct$0.10/1M · 2 providersMai Thinking 1$2.00/1M · 2 providersPhi 4 Mini Instruct$0.075/1M · 1 providerPhi 4 Mini Reasoning$0.075/1M · 1 providerPhi 4 Multimodal Instruct$0.08/1M · 1 providerPhi 2 3B$0.10/1M · 1 providerPhi 4 Reasoning$0.125/1M · 1 provider
Popular comparisons
head-to-head pages featuring Phi 4
Phi 4 vs Gemma 3 12B Listed prices from LiteLLM, OpenRouter and Vercel · last checked Oct 14, 2026
Every weekday
AI moves fast. Here's your debrief.
News, analysis, tools, and more.
For people who build with AI