GPT-4o (Aug '24)
OpenAIGPTProprietary
Description
GPT-4o ('o' for 'omni') is a multimodal AI model that accepts text, audio, image, and video inputs, and generates text, audio, and image outputs. It matches GPT-4 Turbo performance on text and code, with improvements in non-English languages, vision, and audio understanding.
Release Date
2024-08-06
Parameters
—
Context Length
128K
Modalities
image, pdf, text
Capability Radar
50
general
30
coding
50
reasoning
60
scienceest.
50
agents
90
multimodal
Science is estimated from LLM Stats science scores or reasoning when no dedicated science benchmarks are available.
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Multimodal Ranking | 114 | 37.0 | LS |
Benchmark Scores (LLM Stats)
(LLM Stats (zeroeval))Chat
IFEvalGoogle Research (2023)
81.0%SR
Multi-IF
60.9%SR
TAU-bench Retail
60.3%SR
Tau2 Airline
45.5%SR
Multi-Challenge
40.3%SR
Communication
Tau2 Retail
63.4%SR
Tau2 Telecom
23.5%SR
Factuality
SimpleQA
38.2%SR
General
MMLU
85.7%SR
Aider-Polyglot
30.7%SR
Internal API instruction following (hard)
29.2%SR
Aider-Polyglot Edit
18.2%SR
Language
MMMLU
81.4%SR
MMLU-Pro
74.7%SR
COLLIE
61.0%SR
Long Context
ComplexFuncBench
66.5%SR
OpenAI-MRCR: 2 needle 128k
31.9%SR
Math
MathVista
61.4%SR
AIME 2024
13.1%SR
Multimodal
MMMU
72.2%SR
VideoMMMU
61.2%SR
Reasoning
ChartQAMasry et al. (2022)
85.7%SR
CharXiv-D
85.3%SR
GPQANYU + Cohere + Anthropic (2023)
70.1%SR
CharXiv-R
58.8%SR
TAU-bench Airline
42.8%SR
Graphwalks BFS <128k
41.7%SR
Graphwalks parents <128k
35.4%SR
SWE-Bench Verified
33.2%SR
SWE-Lancer
32.6%SR
SWE-Lancer (IC-Diamond subset)
12.4%SR
Humanity's Last Exam
5.3%SR
Vision
AI2D
94.2%SR
DocVQADocVQA (2020)
92.8%SR
EgoSchema
72.2%SR
ActivityNet
61.9%SR
MMMU-Pro
59.9%SR
ERQA
35.2%SR
AA Evaluation Indices
(Artificial Analysis)No AA evaluation data available
LLM Stats Category Scores
(LLM Stats (zeroeval))Image To Text90
Legal80
Finance80
Instruction Following70
Language70
Multimodal70
Physics70
Healthcare70
Biology70
Chemistry70
Vision70
Chat60
Long Context60
Structured Output60
Writing60
Math50
Reasoning50
General50
Communication50
Tool Calling50
Spatial Reasoning40
Factuality40
Frontend Development30
Code30
Pricing
Input Price$2.5 / 1M tokens
Output Price$10 / 1M tokens
Blended Price (3:1)$4.375 / 1M tokens
Cache Read Price$1.25 / 1M tokens
Speed
Tokens/sec165.7
Time to First Token1.84s
Time to Answer1.84s
Provider Price Ranking
Provider Price Ranking
2 providers
Cheapest: OpenAIMost Expensive: Azure
ProviderInputOutput
1OpenAIPRIMARY
$2.5
$10
2Azure
$2.5
$10
Compare pricing across different API providers for this model.