OpenAI Flex Processing: Cheaper AI for Less Urgent Tasks

OpenAI Introduces Flex Processing for Cost-Effective AI Access

Responding to increasing competition from companies like Google in the artificial intelligence sector, OpenAI has announced the availability of Flex processing. This new API option offers reduced costs for utilizing AI models, though it involves trade-offs in speed and reliability.

Lower Costs with Trade-offs

Flex processing is currently in beta for OpenAI’s o3 and o4-mini reasoning models. It is specifically designed for tasks that are not time-critical and do not require constant uptime, such as model testing, data augmentation, and background processing.

The cost reduction is substantial, representing a 50% decrease in API expenses. For the o3 model, Flex processing lowers the price to $5 per million input tokens (approximately 750,000 words) and $20 per million output tokens.

This contrasts with the standard pricing of $10 per million input tokens and $40 per million output tokens for o3. Similarly, o4-mini sees a price reduction to $0.55 per million input tokens and $2.20 per million output tokens, down from $1.10 and $4.40 respectively.

Competitive Landscape and Model Pricing

The introduction of Flex processing arrives as the cost of advanced AI models continues to rise. Furthermore, competitors are releasing more affordable and efficient alternatives.

Just recently, Google launched Gemini 2.5 Flash, a reasoning model that demonstrates comparable or superior performance to DeepSeek’s R1, while offering a lower cost per input token.

Enhanced Security Measures

OpenAI also communicated to its customers that developers in usage tiers 1 through 3 will now be required to complete a new ID verification process to gain access to the o3 model.

These tiers are determined by the total expenditure on OpenAI’s services. Access to reasoning summaries and streaming API functionality for o3, and potentially other models, is also now contingent upon successful verification.

OpenAI has stated that this ID verification is a proactive measure to prevent malicious actors from breaching its usage guidelines and policies.