The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →On March 1, 2023, OpenAI announced an API that let third-party developers build with GPT-3.5 Turbo, the model behind ChatGPT. That opened access to ChatGPT’s underlying technology—not to the ChatGPT application itself. The launch drew attention for its reported low price, broader uses beyond chat, and questions about who might be responsible for AI-generated content.
What the ChatGPT API launch actually opened up
An API, or application programming interface, lets one software product send requests to another and receive results. OpenAI’s announcement gave outside developers a way to use GPT-3.5 Turbo in their own products. It did not give them the ChatGPT website or app as a product to embed or rebrand.
| Option | What it meant in the 2023 launch coverage |
|---|---|
| ChatGPT application | OpenAI’s consumer-facing chat product. |
| ChatGPT API | A developer interface to GPT-3.5 Turbo, which developers could connect to their own applications and workflows. |
IEEE Spectrum reported that GPT-3.5 Turbo also powered ChatGPT and Microsoft’s Bing Chat at the time. The API distinction matters: developers could build a conversational feature, but they could also use the model in software that did not resemble a chat app.
What the launch-era price and performance claims meant
IEEE Spectrum reported a launch-era price of US $0.002 per 1,000 tokens. The article estimated that 1,000 tokens corresponded to roughly 750 English words. That is an approximate conversion, not a fixed rule: token counts vary with the text, and the reported price describes the 2023 launch, not current API pricing.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
#1 Best Overall
The article characterized the price as a substantial decrease from GPT-3. Hassan El Mghari, then a senior developer advocate at Vercel, called the API “10x cheaper and 5x faster” than GPT-3, adding, “It’s extremely rare for a company to release a new version of its API that’s both 10x cheaper and 5x faster.” Those are El Mghari’s attributed comparisons, not independently established benchmarks in the article.
The lower reported price suggested that developers might find it practical to add AI features to products where earlier costs had been a barrier. Whether a particular feature would be affordable depended on its usage; the launch price alone does not establish what an application would cost to operate.
Rank #2
Why developers did not have to build only chatbots
ChatGPT’s popularity made it easy to assume that its API was useful only for conversational interfaces. Chanyeol Choi, CEO and cofounder of Publishd, warned against that assumption: “Due to ChatGPT’s rise in popularity because of its chat format, people tend to have a preconception that ChatGPT API can only be used in this casual format.”
IEEE Spectrum described OpenAI as emphasizing uses beyond casual conversation. An API can provide model output within a larger workflow or product, so the user need not interact with it as a standalone chatbot. Publishd, an AI writing assistant for academics and researchers, was described as using retrieval grounding and being in closed beta at the time of the article.
Rank #3
Context limits and model snapshots
The article also highlighted longer context limits and the ability for developers to pin a model snapshot—a particular model version—to improve consistency. Pinning a snapshot can help keep an application’s behavior from changing unexpectedly when a model is updated; it does not guarantee identical outputs for every request.
Kyle Shannon, Storyvine’s CEO and founder, said that OpenAI’s best dedicated server plans could handle up to 32,000 tokens. That figure is Shannon’s historical description of those plans, not a general statement about every API configuration or a current limit. The article did not establish current context windows or present-day plan terms.
What companies said they were building
IEEE Spectrum named Instacart, Shopify, and Snap as companies announcing API-powered features. Its example for Instacart was helping users decide what they could have for lunch. These examples illustrated that the model could be integrated into an existing service rather than offered only as a separate chat destination.
- Instacart: help for users deciding what to eat.
- Shopify: an announced API-powered feature; the article did not specify further details in the reporting summarized here.
- Snap: an announced API-powered feature; the article did not specify further details in the reporting summarized here.
The legal question raised by generated content
The 2023 article placed the API launch amid a then-pending U.S. Supreme Court challenge involving Section 230. During a hypothetical discussion, Justice Neil Gorsuch said: “Artificial intelligence generates poetry,” and “It generates polemics today that would be content that goes beyond picking, choosing, analyzing, or digesting content. And that is not protected. Let’s assume that’s right.”
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesBest Value
That exchange was not a ruling about ChatGPT, API developers, or the liability of model providers. IEEE Spectrum reported uncertainty about whether developers using generative AI or the providers of the models could be held liable for generated content. The remarks were hypothetical and historical; they do not establish a legal outcome or provide legal advice.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to read the announcement today
The announcement was significant because it made GPT-3.5 Turbo available for integration by third-party developers, and because the reported launch pricing and capabilities could make new kinds of applications viable. The figures and product details above describe the March 2023 launch coverage by IEEE Spectrum. They should not be treated as current API prices, model availability, or limits.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
