OpenAI Launches Flex Processing: Lower API Costs for o3 and o4-mini with Trade-Offs
Image Credit: Jacky Lee
OpenAI, a leading artificial intelligence research organization, has introduced Flex processing, a new beta API option announced on April 17, 2025, aimed at making its advanced AI models more affordable. Flex processing offers halved costs for using OpenAI’s o3 and o4-mini reasoning models but comes with trade-offs like slower response times and occasional resource unavailability. This move responds to growing competition from rivals like Google and DeepSeek, who are aggressively targeting cost-conscious AI developers.
[Read More: ChatGPT Pro vs. Plus: Is OpenAI's $200 Plan Worth the Upgrade?]
What Is Flex Processing?
Flex processing is a budget-friendly API tier designed for developers who prioritize cost savings over speed. Tailored for non-urgent tasks—such as data enrichment, model evaluations, and asynchronous workflows—Flex processing leverages OpenAI’s idle GPU resources to lower operational costs. However, Flex jobs are queued behind standard priority tasks, meaning developers may experience slower response times and periodic unavailability during peak usage.
Flex processing offers a 50% cost reduction compared to standard API rates:
o3: $5 per million input tokens (approx. 750,000 words) and $20 per million output tokens (compared to $10 and $40 standard pricing).
o4-mini: $0.55 per million input tokens and $2.20 per million output tokens (compared to $1.10 and $4.40 standard pricing).
Flex processing is ideal for “non-production” scenarios and is accessible to developers who meet OpenAI’s newly introduced ID verification requirements.
[Read More: AI Breakthrough: OpenAI’s o1 Model Poised to Surpass Human Intelligence]
Competitive Context
The launch of Flex processing comes amid an increasingly competitive AI landscape. Just a day before, Google unveiled Gemini 2.5 Flash, a reasoning model offering strong performance at competitive pricing, while DeepSeek's R1—introduced earlier in January—continues to disrupt with ultra-low pricing ($0.14 per million input tokens and $2.19 per million output tokens).
Although OpenAI’s Flex pricing narrows the gap, it remains more expensive than DeepSeek’s offerings. For example, o4-mini’s Flex input cost of $0.55 is closer to Gemini 2.5 Flash’s pricing but still significantly higher than DeepSeek R1’s. This strategic adjustment signals OpenAI’s recognition of growing pressure to deliver cost-efficient, capable AI services.
ID Verification Requirement
OpenAI has simultaneously introduced an ID verification process for developers in usage tiers 1–3, based on their spending levels. Verified identification is now mandatory to access Flex processing for the o3 model, reasoning summaries, and streaming API features.
OpenAI states that this measure enhances platform security by preventing malicious actors from abusing the system—a move consistent with broader industry efforts to tighten user controls. However, it may create barriers for some smaller developers, especially those concerned about privacy or facing regional documentation challenges.
[Read More: OpenAI’s ChatGPT Introduces Studio Ghibli-Style Image Feature]
Implications for Developers and Businesses
Flex processing offers a compelling opportunity for startups, research teams, and budget-conscious businesses. By cutting API costs in half, developers can leverage advanced models like o3 and o4-mini for non-time-sensitive projects such as dataset enrichment, model benchmarking, and background analytics, all without bearing premium fees.
However, for real-time applications—such as customer-facing chatbots or interactive tools—Flex processing’s slower response times and occasional queuing delays make it less ideal. Meanwhile, the ID verification requirement could add a layer of friction, particularly for developers operating in regions with strict privacy laws or complex ID frameworks.
Source: Tech Crunch, Winbuzzer, Gadgets360, OpenAI
