Intelligence des modèles d'IA

Phi-4-mini-instruct

microsoft/phi-4-mini-instruct

Par Microsoft · famille: phi · sorti 2024-12-11 · fin de connaissance: 2023-10

$0.080
Entrée / 1M jetons
$0.350
Sortie / 1M jetons
128K
Fenêtre de contexte
128K
Sortie max

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

Capacités

✓ Tool calling✓ Raisonnement✓ Sortie structurée✗ Pièces jointes✓ Poids ouverts✓ Contrôle de température
Modalités: entrée text · sortie 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.

Coding76
  • Tool calling40/40
  • Structured output20/20
  • Reasoning10/10
  • Context window (100K → 1M)2/20
  • Provider availability4/10
Agents94
  • Tool calling35/35
  • Structured output25/25
  • Reasoning15/15
  • Output token limit15/15
  • Provider availability4/10
JSON / structured output99
  • Structured output / JSON mode50/50
  • Tool calling20/20
  • Temperature control10/10
  • Price-friendly for high-volume19/20
Cost efficiency71
  • Headline price (log-scaled)71/95
  • Has prompt-cache pricing0/5
Long context45
  • Context window (100K → 2M)35/90
  • Has published price for full window10/10
Production-readiness74
  • Number of independent providers20/40
  • Has published per-token price20/20
  • Context window ≥ 8K15/15
  • No data inconsistencies across providers4/10
  • 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
$0.57
< $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
$1.15
< $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
$0.34
< $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
$0.99
< $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
$1.17
< $0.01 per request
12K input tokens (long-running tool history) and a 600-token tool-call decision. Cost per agent step.

Détail des tarifs

Tarif recommandé de wandb · microsoft/Phi-4-mini-instruct

$0.080
Entrée
$0.350
Sortie

Fournisseur le moins cher : nvidia · Unknown entrée + Unknown sortie

Disponible chez 4 fournisseurs

FournisseurID modèle fournisseurEntrée / 1MSortie / 1MContextePublié le
Weights & Biases
wandb
microsoft/Phi-4-mini-instruct$0.080$0.350128K2024-12-11
Kilo Gateway
kilo
microsoft/phi-4-mini-instruct$0.080$0.350128K2025-10-17
Nvidia
nvidia
microsoft/phi-4-mini-instructUnknownUnknown131K2024-12-01
GitHub Models
github-models
microsoft/phi-4-mini-instructUnknownUnknown128K2024-12-11

Incohérences de données entre fournisseurs

  • context_window varies: 128000, 131072
  • release_date varies (span 320d): 2024-12-01, 2024-12-11, 2025-10-17
  • modalities varies across offerings

Les fournisseurs rapportent des valeurs différentes pour ce modèle. Les infos clés ci-dessus utilisent un fournisseur représentatif ; voir le tableau pour le détail par fournisseur.

Frequently asked questions

How much does Phi-4-mini-instruct cost?

Phi-4-mini-instruct costs $0.080 per 1M input tokens and $0.350 per 1M output tokens, sourced from wandb. 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-mini-instruct?

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

Does Phi-4-mini-instruct support tool calling?

Yes. Phi-4-mini-instruct supports tool calling (function calling). This makes it suitable for production agent and automation workloads where the model has to invoke external functions reliably.

Does Phi-4-mini-instruct support structured output / JSON mode?

Yes. Phi-4-mini-instruct supports structured output / JSON-schema-constrained decoding. This makes it suitable for production agent and automation workloads where the model has to invoke external functions reliably.

Can Phi-4-mini-instruct accept image input?

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

Is Phi-4-mini-instruct open-weight?

Yes. Phi-4-mini-instruct'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-mini-instruct?

If Phi-4-mini-instruct doesn't fit, consider Phi-4, 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

Dernière mise à jour :

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.