Phi-4 — Especificações
Compiled by Mustafa Ihsan from the vendor’s published documentation · Last updated
| Desenvolvedor | Microsoft |
|---|---|
| Tipo | LLM (densa) |
| Modalidade | Texto → Texto |
| Parâmetros | 14B |
| Janela de contexto | 16K |
| Licença | MIT (aberta) |
| Pesos abertos | Sim |
| Lançado | 2025 |
| Preço da entrada | $0.07 /1M |
| Preço da saída | $0.14 /1M |
| Provedores de API | Azure, OpenRouter, Ollama |
Execute-o localmente
| VRAM (4 bits) | ~9 GB |
|---|---|
| GPU mínima | RTX 4070 12 GB / RTX 3060 12 GB |
What is Phi-4?
Phi-4 is Microsoft’s compact 14B reasoning model, MIT-licensed, which punches well above
its size on mathematics and logic. It needs about 9 GB of VRAM at 4-bit, running comfortably
on an RTX 4070 or RTX 3060 12GB, and costs $0.07 in / $0.14 out per million tokens
hosted.
The Phi line’s whole thesis is that curated, textbook-quality training data beats raw
scale for reasoning tasks, and Phi-4 is the clearest evidence for it — a 14B model competing
on maths and logic benchmarks with models several times larger. The cost of that focus is the
16K context window, by far the narrowest in this database and a hard limit for any workload
involving documents, long conversations or retrieval. Read it as a specialist: excellent for
structured reasoning over short inputs — maths tutoring, logic and code puzzles, deterministic
extraction from small payloads — and the wrong tool the moment your prompt grows. If you need
Phi-4’s reasoning with room to work, a 128K-context model in the same hardware bracket such
as Qwen3 14B is the better trade.
Phi-4 pricing: API cost per 1M tokens
| Entrada (por 1 milhão de tokens) | $0.0700 |
|---|---|
| Saída (por 1 milhão de tokens) | $0.140 |
| Razão saída/entrada | 2× |
| Misturada (4:1 entrada:saída) | $0.0840 por 1 milhão de tokens |
What Phi-4 costs per month
Gasto mensal real com uma proporção de entrada para saída de 4:1 — a razão efetivamente gerada por cargas de trabalho típicas de chat ou RAG.
| Carga de trabalho | Tokens/mês | Custo por mês |
|---|---|---|
| Projeto paralelo | 1 milhão de tokens de entrada / 0,25 milhão de tokens de saída | $0.11 |
| Equipe pequena | 20 milhões de tokens de entrada / 5 milhões de tokens de saída | $2.10 |
| Produção | 200 milhões de tokens de entrada / 50 milhões de tokens de saída | $21 |
Calcule seus próprios números na Calculadora de custos de API de IA.
Cheaper alternatives to Phi-4
| Modelo | Custo médio ponderado por US$ 1 milhão | Você economiza |
|---|---|---|
| Mistral 7B aberta | $0.0220 | 74% cheaper |
| Llama 3.1 8B aberta | $0.0220 | 74% cheaper |
| Mistral NeMo 12B aberta | $0.0240 | 71% mais barato |
Hospedar localmente ou pagar pela API?
Phi-4 is open-weight, so you can run it yourself. It needs ~9 GB of VRAM at 4-bit (RTX 4070 12GB / RTX 3060 12GB). Self-hosting only beats the API once your volume is high enough to keep that hardware busy — the calculadora de autohospedagem versus API calcula o ponto de equilíbrio para seu volume de tokens.
Perguntas frequentes
How much does Phi-4 cost per 1M tokens?
Phi-4 costs $0.0700 per 1M input tokens and $0.140 per 1M output tokens. At a typical 4:1 input-to-output mix that blends to about $0.0840 per 1M tokens.
How much does Phi-4 cost per month?
A small-team workload of 20M input and 5M output tokens a month costs about $2.10 on Phi-4. A side project (1M in / 0.25M out) costs roughly $0.11.
What is a cheaper alternative to Phi-4?
Mistral 7B is the strongest cheaper option in our database at $0.0220 per 1M blended — about 74% less than Phi-4. It is also open-weight, so self-hosting is an option.
Can I run Phi-4 locally?
Yes. Phi-4 is open-weight and needs about ~9 GB of VRAM at 4-bit quantisation (RTX 4070 12GB / RTX 3060 12GB).
Why does Phi-4 charge more for output than input?
Output tokens are generated one at a time and cannot be batched the way a prompt can, so they cost the provider more to serve. Phi-4 charges 2× more for output, which is why prompt-heavy workloads are far cheaper to run than generation-heavy ones.
Os preços correspondem às tarifas listadas oficialmente para a API principal do modelo e são revisados conforme os provedores os atualizam. Descontos por volume, processamento em lote ou entradas em cache não estão incluídos. Compare todos os modelos lado a lado na Banco de dados de modelos de IA ou o Ranking de LLMs.