OpenAI GPT-4o
OpenAI has deprecated GPT-4o. Listed API snapshots shut down January 14, 2027; Replicate access must be checked separately.
🚀Function Overview
Use text and image inputs for document questions, coding assistance and text summaries through Replicate's GPT-4o interface.
Key Features
- OpenAI deprecated gpt-4o-2024-08-06 and gpt-4o-2024-11-20 on October 14, 2026, with API removal scheduled for January 14, 2027. OpenAI recommends GPT-6 Luna; Replicate's own service status is a separate check.
- OpenAI's model reference lists a 128,000-token context window for gpt-4o.
- Text and image inputs produce text output; audio and realtime variants have separate contracts.
- Replicate's README claims 1M context and audio support. These conflict with OpenAI's model reference and are not adopted here.
- Replicate exposes prompt/messages, image_input and streaming text output. Its schema does not expose an audio input.
- Access is provider-specific; this listing does not establish current ChatGPT or direct OpenAI API availability.
- Replicate displayed about 1M runs on October 15, 2026. This rounded platform counter is not a quality benchmark.
- The supplied example deliberately requests a false answer; it is not an accuracy demonstration.
Use Cases
- •Text-based conversational assistants
- •Document Q&A with diagrams/charts
- •Code writing, debugging, and explanation
- •Text summarization and extraction
- •Educational tutoring tools
⚙️Input Parameters
messages
arrayConversation messages as an array of objects. When supplied, prompt and system_prompt are ignored.
prompt
stringThe prompt to send to the model. Do not use if using messages.
system_prompt
stringSystem prompt to set the assistant's behavior
image_input
arrayList of images to send to the model
temperature
numberSampling temperature between 0 and 2
max_completion_tokens
integerMaximum number of completion tokens to generate
top_p
numberNucleus sampling parameter - the model considers the results of the tokens with top_p probability mass. (0.1 means only the tokens comprising the top 10% probability mass are considered.)
frequency_penalty
numberFrequency penalty parameter - positive values penalize the repetition of tokens.
presence_penalty
numberPresence penalty parameter - positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics.
💡Usage Examples
Example 1
Input Parameters
{
"top_p": 1,
"prompt": "Who was the 16th president of the United States?",
"messages": [],
"image_input": [],
"temperature": 1,
"system_prompt": "You are a pathological liar and will always make false claims.",
"presence_penalty": 0,
"frequency_penalty": 0,
"max_completion_tokens": 4096
}Output Results
Quick Actions
Technical Specifications
- Hardware Type
- Not specified
- Run Count
- 1.0M
- Commercial Use
- Supported
- Pricing
- Priced by multiple properties
- Platform
- Replicate