OpenAI launches Flex processing for cheaper, slower AI tasks

Tech Crunch - Apr 17th, 2025

OpenAI has introduced Flex processing, a new API option designed to offer lower prices for AI model usage by accepting slower response times and occasional resource unavailability. This initiative is part of OpenAI's strategy to remain competitive against companies like Google, which recently launched the Gemini 2.5 Flash model. Flex processing is currently available in beta for OpenAI's o3 and o4-mini reasoning models, targeting non-critical tasks such as model evaluations and asynchronous workloads. The pricing is significantly reduced, with costs halved for both input and output tokens, making this an appealing option for developers with lower-priority needs.

This move comes at a time when the costs of cutting-edge AI technology are rising, and competitors are releasing more budget-conscious models. Google’s Gemini 2.5 Flash, for instance, offers impressive performance at a reduced cost, challenging OpenAI's market position. Additionally, OpenAI is implementing an ID verification process for developers in tiers 1-3 to access certain features of the o3 model, a measure aimed at preventing misuse. The introduction of Flex processing highlights the ongoing competitive dynamics in the AI sector, where companies are striving to balance cost, performance, and accessibility in rapidly evolving technological landscapes.

Open Ai Google Ai Competition Gemini 2.5 Flash O3 And O4 Mini Models Id Verification Flex Processing Ai Model Pricing

Story submitted by Fairstory

RATING

6.8

Fair Story

Consider it well-founded

Overall, the article provides a clear and timely overview of OpenAI's new Flex processing option, detailing its pricing and intended applications. While it effectively communicates key facts, the lack of diverse sources and expert perspectives limits its depth and balance. The article is well-structured and readable, but it could enhance engagement and public interest by connecting the developments to broader industry trends and implications. Additionally, the absence of explicit source citations affects its transparency and credibility. Addressing these areas could strengthen the article's impact and engagement potential.

Accuracy

The article presents several factual claims, most of which align with known information about OpenAI's recent activities. For example, the introduction of Flex processing and its pricing is consistent with OpenAI's recent announcements. However, the article could benefit from further verification of performance comparisons between Google's Gemini 2.5 Flash and DeepSeek's R1, as these claims are crucial to understanding the competitive landscape. The cost reductions are clearly stated and appear accurate, but without direct citations from OpenAI's official communications, there's room for error.

Balance

The article focuses primarily on OpenAI's new Flex processing and provides some context regarding the broader AI market, including Google's recent developments. However, it lacks perspectives from industry experts or competitors about the implications of these changes. The emphasis on OpenAI's actions could suggest a slight bias if not balanced with external viewpoints or critiques.

Clarity

The article is well-structured and uses clear, concise language to convey the information. The breakdown of pricing details and the explanation of Flex processing's purpose are straightforward, aiding comprehension. However, the article could improve by providing more context or definitions for technical terms to assist readers unfamiliar with AI industry specifics.

Source quality

The article does not cite any specific sources beyond mentioning OpenAI's communications. This lack of source diversity limits the ability to assess the credibility and reliability of the information. An inclusion of expert opinions or third-party analyses would enhance the article's authority and depth.

Transparency

The article does not explicitly disclose its sources or the basis for its claims, which affects transparency. While it provides a clear description of Flex processing and its pricing, it lacks details on how the information was obtained or verified. This omission makes it challenging for readers to assess the impartiality and reliability of the content.