OpenAI has launched Flex processing as a more affordable option for developers using its o3 and o4-mini models.
This new feature allows users to cut costs in half, that’s if they are willing to settle for slower processing times and limited availability.
Flex is now in beta, and it targets tasks that do not need real-time results. These include internal evaluations, data enrichment, and background jobs.
By offering a low-cost option, OpenAI plans to support developers working on less time-sensitive projects.
What Is Flex Processing?
Flex processing gives users access to OpenAI’s models at reduced rates. However, the trade-off is speed.
Tasks may take longer to complete, and sometimes, access may be delayed if demand is high.
Although this would cause frustration in major software projects, flex processing is ideal for simpler things.
These are activities like background tasks, non-production workloads, research and testing, and asynchronous processing
How Much Can You Save?
Flex cuts prices by 50%. Here’s what the pricing structure looks like:
| Model | Plan Type | Input Cost (Per Million Tokens) | Output Cost (Per Million Tokens) |
| o3 | Standard | $10.00 | $40.00 |
| o3 | Flex | $5.00 | $20.00 |
| o4-mini | Standard | $1.10 | $4.40 |
| o4-mini | Flex | $0.55 | $2.20 |
Also read: OpenAI Launches o1-Pro, Its Most Expensive AI Model
The Flex Timing
The timing is strategic. AI running costs continue to increase, and competitors like Google are rolling out lower-cost models, including Gemini 2.5 Flash (strong performance at reduced input costs).
OpenAI’s Flex release is a direct response to the competition. It gives developers more control over how they spend, without sacrificing model quality for routine or internal tasks.
Some Limitations to Consider
While Flex helps cut costs, it comes with a few downsides. It has slower response times, limited resource availability, and is unsuitable for production systems
Therefore, flex works best for low-priority jobs. Developers should not rely on it for real-time apps or user-facing tools.
New ID Verification Requirement
OpenAI is also tightening access to its models. Developers in usage tiers 1–3 must now complete an ID verification process to use o3 and related features.
OpenAI says this step is necessary to prevent misuse and protect its platform. The verification process applies to users based on their spending tier.
