What are the limits for the newly announced o3 and o4-mini
- User
What are the limits for the newly announced o3 and o4-mini
- Assistant
OpenAI's newly announced o3 and o4-mini models come with several types of limits, ranging from usage quotas to technical specifications. Here's a comprehensive breakdown of these limitations:
Usage Limits by Subscription Tier
ChatGPT Plus, Team, and Enterprise Users
- o3: Limited to 50 messages per week[1][3][12]
- o4-mini: 150 messages per day[1][3][12][15]
- o4-mini-high: 50 messages per day[1][3][12][15]
ChatGPT Pro Users ($200/month)
- "Near unlimited" access to o3, o4-mini, and o4-mini-high models[1][3][12]
- Still subject to OpenAI's terms of use, which prohibit:
- Automated or programmatic data extraction
- Sharing account credentials
- Reselling access or powering third-party services[1][3]
Free Users
- Access only to o4-mini through the "Think" or "Reason" button in the interface[1][12][15]
- No access to o3 or o4-mini-high[1][12]
Technical Specifications and Limits
Context Window
- o3: 200,000 tokens maximum input[14][17]
- o4-mini: 200,000 tokens maximum input[14][6]
Maximum Output
- o3: Up to 100,000 tokens per response[14]
- o4-mini: Up to 100,000 tokens per response[14][6]
Practical Implementation Limits
Some users have reported that the actual token limits in the chat interface are lower than advertised:
- Approximately 60-65K tokens per message in the web application[10][11]
- Around 30K tokens per message in the macOS application[10][11]
API Pricing and Limits
o3 API Pricing
- Input: $10.00 per million tokens[5][7]
- Cached input: $2.50 per million tokens[5]
- Output: $40.00 per million tokens[5][7]
o4-mini API Pricing
- Input: $1.10 per million tokens[5][6][7][13]
- Cached input: $0.275 per million tokens[5]
- Output: $4.40 per million tokens[5][6][7][13]
Usage Limit Reset Schedule
For the weekly o3 limit:
- The 50-message limit resets every seven days from when you send your first message[3]
- For example, if you start using o3 on the 12th of the month, your limit resets on the 19th at 00:00 UTC[3]
For daily limits (o4-mini and o4-mini-high):
- Limits reset daily, though the exact reset time is not specified in the search results
You can check when your usage limits will reset by highlighting the model name in the model picker dropdown menu in ChatGPT[3]. Currently, there is no way to check how many messages you have remaining within your usage budget[3].
Verification of Model Usage
To verify which model was used for a response:
- Check at the top of the ChatGPT response
- You can also highlight the regenerate button under any response to see which model was used[3]
These limits reflect OpenAI's attempt to balance access to these powerful new reasoning models while managing computational resources and preventing misuse.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.