MI
Phi 4
Microsoft
Microsoft model product
Phi 4 multimodal
0.
Updated Aug 12, 2026. Default version: Phi-4-multimodal-instruct
LLMBoard score7.4Phi-4-multimodal-instruct
Coverage0%12 benchmark families
Context window128KTokens
Official input priceN/AOfficial price unavailable
Phi 4 multimodal Specifications
Technical details for the model's default version.
- Version
- Phi-4-multimodal-instruct
- Released
- Feb 1, 2025
- Knowledge cutoff
- Jun 1, 2024
- Parameters
- 5.6B
- Context window
- 128K
- Max output
- 128K
- Inputs
- image, text
- Outputs
- text
- Open weights
- No
- License
- MIT
Phi 4 multimodal Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
No capability profile
No version of this model currently has enough benchmark coverage for a score.
Phi 4 multimodal Benchmark Results
Benchmark scores for Phi-4-multimodal-instruct.
| InterGPS | 48.6% | 01 | 2 | 100.0% | C | |
| ScienceQA Visual | 97.5% | 01 | 1 | 100.0% | C | |
| POPE | 85.6% | 02 | 2 | 0.0% | C | |
| MMBench | 86.7% | 03 | 9 | 75.0% | C | |
| InfoVQA | 72.7% | 05 | 9 | 50.0% | C | |
| TextVQA | 75.6% | 10 | 15 | 35.7% | C | |
| BLINK | 61.3% | 13 | 13 | 0.0% | C | |
| DocVQA | 93.2% | 14 | 26 | 48.0% | C | |
| OCRBench | 84.4% | 16 | 22 | 28.6% | C | |
| Video-MME | 55.0% | 17 | 17 | 0.0% | C | |
| ChartQA | 81.4% | 19 | 24 | 21.7% | C | |
| MathVista | 62.4% | 25 | 39 | 36.8% | C | |
| AI2D | 82.3% | 27 | 32 | 16.1% | C | |
| MMMU | 55.1% | 52 | 63 | 17.7% | C | |
| MMMU-Pro | 38.5% | 63 | 66 | 4.6% | C |
Phi 4 multimodal Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
Phi 4 multimodal Runtime Performance
Provider-specific output speed and catalog latency for Phi-4-multimodal-instruct. Runtime does not affect the capability score.
| DeepInfra | 25 tok/s | 0.5 s | 128K | 128K |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Phi 4 multimodal Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- N/A
- Official provider
- N/A
- Lowest third-party
- From $0.07 input, $0.11 output per 1M via NanoGPT
- Tracked offerings
- 1
| NanoGPT | phi-4-multimodal-instruct | global | $0.07 | $0.11 | 128K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Phi 4 multimodal Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| Phi-4-multimodal-instruct | 7.4 | 5.6B | 128K | 128K | No | MIT |
What is Phi 4 multimodal?
Key information about Phi 4 multimodal and its available data.
0. It processes text, image, and audio inputs to generate text outputs, supporting a 128K token context length. Enhanced via SFT, DPO, and RLHF for instruction following and safety.
Data as of 2026-08-11.
Phi 4 multimodal vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Phi 4 multimodal
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
MI
Phi 3.5 MoE
Microsoft
MI
Phi 4 Mini Reasoning
Microsoft
MI
Phi 3.5 vision
Microsoft
MI
Phi 3.5 mini
Microsoft
MI
Phi 4 Mini
Microsoft
FAQ
Common questions about Phi 4 multimodal.
When was Phi 4 multimodal released?
Phi 4 multimodal's default version was released on Feb 1, 2025.
How much does Phi 4 multimodal cost?
No official standard PAYG price is currently available for Phi 4 multimodal. The lowest tracked third-party offer starts at $0.07 input and $0.11 output via NanoGPT.
Who created Phi 4 multimodal?
Phi 4 multimodal was created by Microsoft.
What is the context window for Phi 4 multimodal?
The default version has a 128K token context window.
Is Phi 4 multimodal open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer Phi 4 multimodal?
1 provider offerings are linked to the default version.
What models should I compare Phi 4 multimodal with?
Nearby ranked alternatives include Qwen2.5 14B, GPT-4o-mini, Mistral Small 3 24B.