Open-weight instruction model for adaptable chat and self-hosted production workloads
[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed. At 14 billion...
Context
128K
128,000 tokens
Max output
4K
4,096 tokens
Input
$0.13
per 1M tokens
Output
$0.50
per 1M tokens
Providers
3
3 listings
Intelligence
5.9
Intelligence Index
Pricing
USD per 1M tokens · listed by Azure
| Token type | Standard |
|---|---|
| Input | $0.13 |
| Output | $0.50 |
Lowest input price across providers: $0.06.
Capabilities
ReasoningNo
Tool callNo
Structured outputYes
TemperatureYes
AttachmentsNo
Interleaved reasoningNo
Open weightsYes
ModeratedNo
Input
Text
Output
Text
Benchmarks
Published scores and Design Arena Elo
- Intelligence5.9#198 of 208
- CodingNot scored—
- AgenticNot scored—
- ArenaNot scored—
Rankings
Position on each OneAtlas leaderboard
Providers · 3
3 listings
| Provider | Model ID | Context | Input limit | Output | Input $ | Output $ | Reasoning $ | Cache read | Cache write | Audio in | Audio out | Reasoning | Tools | Structured | Temp. | Updated |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| phi-4 | 128K | — | 4K | $0.13 | $0.50 | — | — | — | — | — | No | No | Unknown | Yes | Dec 11, 2024 | |
| phi-4 | 128K | — | 4K | $0.13 | $0.50 | — | — | — | — | — | No | No | Unknown | Yes | Dec 11, 2024 | |
| microsoft/phi-4 | 16K | — | 15K | $0.06 | $0.11 | — | — | — | — | — | No | No | Yes | Yes | Jan 10, 2025 |