OpenAI Launches Decisions API for Faster Text and Image Classification
OpenAI has introduced a Decisions API that classifies text and images roughly ten times faster than its Responses API, returning yes/no probabilities, category picks, or scale ratings at $0.10 per million input tokens. The company also reduced its paid API tiers from five to three.
OpenAI has released a new Decisions API designed to simplify complex evaluations by reducing them to binary choices, category selections, or scale ratings. According to the company, the tool classifies text and images about ten times faster than its existing Responses API, offering a streamlined option for developers who need quick, structured outputs rather than open-ended generation.
The API returns yes/no probabilities, allows users to pick one category from a set, or provides ratings on a scale. This makes it suited for tasks such as content moderation, sentiment analysis, survey coding, and any workflow where a clear decision is more useful than a lengthy text response. Pricing is set at $0.10 per million input tokens, positioning it as a low-cost alternative for high-volume classification jobs.
Alongside the launch, OpenAI has restructured its paid API tiers, cutting the number from five down to three. The change is intended to simplify access for developers and businesses that rely on the company’s models for production applications. While the exact composition of the new tiers was not detailed, the reduction signals a broader effort to make the platform easier to navigate as competition in the AI API market intensifies.
The Decisions API reflects a growing demand for specialized AI tools that do one thing well. Instead of prompting a large language model to reason through a problem and produce a paragraph of analysis, developers can now send text or images and receive a direct answer. This can reduce latency, lower costs, and make it easier to integrate AI into existing decision pipelines.
For sectors such as media, education, and customer service, the ability to quickly classify content could speed up moderation and personalization. In healthcare and research, binary or categorical outputs might assist with triaging information, though such uses would still require human oversight. The API’s speed advantage over the Responses API suggests it is optimized for throughput rather than deep reasoning.
OpenAI’s move comes as other providers offer similar classification endpoints, and the company appears to be betting that simplicity and price will attract developers. By limiting the output to a decision, the API also reduces the risk of ambiguous or off-topic responses, which can be a problem when using generative models for structured tasks.
The reduction in paid tiers may also affect existing customers who need to adjust their subscriptions. OpenAI has not announced any changes to the underlying models or to the free tier, but the consolidation could push some users toward higher-volume plans. The company has been steadily expanding its enterprise offerings, and a simpler pricing structure may help it compete for business clients who prefer predictable costs.
Overall, the Decisions API represents a shift toward practical, task-specific AI services. By focusing on speed and low cost, OpenAI is making it easier for developers to embed classification into their products without building custom models. The success of the API will likely depend on how well it handles nuanced cases where a simple yes or no is not enough, but for many routine decisions, it offers a faster path.
6
