跳转到主要内容

Phi-4 Multimodal Instruct

MicrosoftPhi开源权重MIT · 商用许可

描述

Phi-4-multimodal-instruct is a lightweight (5.57B parameters) open multimodal foundation model that leverages research and datasets from Phi-3.5 and 4.0. 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.

发布日期
2025-02-26
参数规模
5.6B
上下文长度
—
支持模态
image, text

能力雷达图

18
general
13
coding
32
reasoning
23
science
26
agents
85
multimodal

排行榜排名

领域#排名分数来源
代码能力榜563
14.0
AA
通用能力榜584
20.0
AA
多模态榜141
23.0
LS
科学能力607
16.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Math

MathVista62.4%自报
InterGPS48.6%自报

Multimodal

MMMU55.1%自报
Video-MME55.0%自报

Reasoning

ChartQAMasry et al. (2022)81.4%自报

Vision

ScienceQA Visual97.5%自报
DocVQADocVQA (2020)93.2%自报
MMBench86.7%自报
POPE85.6%自报
OCRBench84.4%自报
AI2D82.3%自报
TextVQA75.6%自报
InfoVQA72.7%自报
BLINK61.3%自报
MMMU-Pro38.5%自报

AA 评测指数

(Artificial Analysis)
Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))
69.3
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))
48.5
Gpqa(NYU + Cohere + Anthropic (2023))
31.5
Livecodebench(UC Berkeley + MIT + Cornell (2024))
13.1
Aime(MAA (Mathematical Association of America))
9.3
Intelligence Index(Artificial Analysis)
5.8
Hle(Center for AI Safety + Scale AI (2025))
5.0

LLM Stats 分类评分

(LLM Stats (zeroeval))
Safety
90
Image To Text
80
Multimodal
70
Reasoning
70
General
70
Vision
70
Math
60
Spatial Reasoning
60
Healthcare
60
3d
60

定价

输入价格免费
输出价格免费
混合价格(3:1)免费

速度

Tokens/秒17.2
首Token延迟0.35s
首回答延迟0.35s

供应商价格排行

供应商价格排行

2 个供应商

最便宜: Azure Cognitive Services最贵: Azure
供应商输入输出
1Azure Cognitive Services最便宜
$0.08
$0.32
2Azure
$0.08
$0.32

比较该模型在不同 API 供应商之间的定价。

外部链接