Inteligência em modelos de IA

Phi-4

microsoft/phi-4

Por Microsoft · família: phi · lançado 2024-12-11 · data de conhecimento: 2023-10

$0.060
Entrada / 1M tokens
$0.140
Saída / 1M tokens
128K
Janela de contexto
4K
Saída máxima

Prices in USD per 1M tokens. Unknown means the provider does not publish per-token pricing.

Capacidades

✗ Tool calling✗ Raciocínio? Saída estruturada✗ Anexos✓ Pesos abertos✓ Controle de temperatura
Modalidades: entrada text · saída text

Model fit scores

0–100 · higher is better

These scores reward declared capabilities, context size, price and provider availability — they are not benchmark results. Use them as a directional signal alongside your own evaluation.

Coding6
  • Tool calling0/40
  • Structured output0/20
  • Reasoning0/10
  • Context window (100K → 1M)2/20
  • Provider availability4/10
Agents4
  • Tool calling0/35
  • Structured output0/25
  • Reasoning0/15
  • Output token limit0/15
  • Provider availability4/10
JSON / structured output30
  • Structured output / JSON mode0/50
  • Tool calling0/20
  • Temperature control10/10
  • Price-friendly for high-volume20/20
Cost efficiency79
  • Headline price (log-scaled)79/95
  • Has prompt-cache pricing0/5
Long context45
  • Context window (100K → 2M)35/90
  • Has published price for full window10/10
Production-readiness78
  • Number of independent providers20/40
  • Has published per-token price20/20
  • Context window ≥ 8K15/15
  • No data inconsistencies across providers8/10
  • Official model (not derivative)15/15

Cost Efficiency Index

Open full calculator →

Estimated cost using the recommended provider's headline rate. Each scenario fixes average input/output tokens — the assumptions are shown in the third column.

ScenarioCostAssumption
RAG answer
per 1,000 RAG answers
$0.37
< $0.01 per request
5K input tokens (query + 4 retrieved chunks of ~1K each) and a 500-token answer. Typical SaaS knowledge-base bot.
Support ticket triage
per 10,000 tickets
$0.74
< $0.01 per request
1K input tokens (ticket body + system prompt) and a 100-token JSON classification reply. High-volume customer support.
Data extraction
per 1,000 documents
$0.19
< $0.01 per request
2K input tokens (a single document page) and a 500-token JSON extraction. ETL / invoice / form pipelines.
Code review
per 1,000 PRs
$0.62
< $0.01 per request
8K input tokens (diff + surrounding files) and a 1K-token review comment. PR-bot workloads.
Agent step
per 1,000 steps
$0.80
< $0.01 per request
12K input tokens (long-running tool history) and a 600-token tool-call decision. Cost per agent step.

Detalhes de preço

Preço recomendado de kilo · microsoft/phi-4

$0.060
Entrada
$0.140
Saída

Provedor mais barato: github-models · Unknown entrada + Unknown saída

Disponível em 4 provedores

ProvedorID do modelo do provedorEntrada / 1MSaída / 1MContextoLançado
Azure
azure
phi-4$0.125$0.500128K2024-12-11
Kilo Gateway
kilo
microsoft/phi-4$0.060$0.14016K2024-12-11
Azure Cognitive Services
azure-cognitive-services
phi-4$0.125$0.500128K2024-12-11
GitHub Models
github-models
microsoft/phi-4UnknownUnknown16K2024-12-11

Inconsistências de dados entre provedores

  • context_window varies: 128000, 16000, 16384

Os provedores reportam valores diferentes para este modelo. Os dados rápidos acima usam um provedor representativo; consulte a tabela para detalhes por provedor.

Frequently asked questions

How much does Phi-4 cost?

Phi-4 costs $0.060 per 1M input tokens and $0.140 per 1M output tokens, sourced from kilo. Cache reads, audio tokens and >200K-context tiers (where applicable) are listed in the Pricing detail block above.

What is the context window of Phi-4?

Phi-4 has a context window of 128K tokens, with a max output of 4K tokens per reply. This is the total combined size of prompt + completion.

Does Phi-4 support tool calling?

No. Phi-4 does not support tool calling (function calling). If your workflow requires it, look at the /capabilities/tool-calling list for alternatives.

Does Phi-4 support structured output / JSON mode?

Support for structured output / JSON-schema-constrained decoding is not reported for Phi-4 in our data source. Verify with Microsoft's official documentation before relying on it in production.

Can Phi-4 accept image input?

No. Phi-4 only accepts text as input. If you need image input, see our /capabilities/vision list for current vision-capable models.

Is Phi-4 open-weight?

Yes. Phi-4's weights are publicly available, so you can self-host or fine-tune. Note that open weights ≠ open source — the training data and code are typically not released.

What are the best alternatives to Phi-4?

If Phi-4 doesn't fit, consider Phi-4-mini-instruct, WizardLM-2 8x22B, Phi-3-medium-instruct (128k). Each one targets the same use case — see the Related links below for direct head-to-head pages.

Where does this data come from?

All numbers come from the public models.dev API and are normalised into a single canonical model record. We re-pull daily and write any changes (price, context, capability) to the /changelog page.

Explore more

More Microsoft models

Capability lists this model is in

See also

Última atualização:

Data is sourced from models.dev and normalized for comparison. Prices and capabilities may change. Always verify critical production decisions with the provider's official documentation.