OpenAI adds o1 to its API, but only for certain developers to start, announces new versions of GPT-4o and GPT-4o mini as part of its Realtime API, and more
OpenAI is bringing o1, its “reasoning” AI model, to its API — but only for certain developers, to start.
TechCrunchKyle Wiggers
Context & Ripple Effects
OpenAI had previously taken GPT-4’s API from limited availability to broad access for paying developers through its GPT-4 API general-availability rollout. This update applies that API-distribution playbook to o1, but begins with a more selective gate.
Selected developers can begin building with o1 through OpenAI’s API, while developers outside the initial cohort remain unable to use it there.
Developers using the Realtime API receive updated GPT-4o and GPT-4o mini options, expanding the models available for real-time application workflows.
Second-order effects
Teams evaluating o1 will need to choose between its reasoning-oriented behavior and the cost, speed, and performance trade-offs identified in earlier coverage, rather than treating it as a simple GPT-4o replacement.
If reasoning models continue entering APIs in staged tiers, model selection becomes a product and operations decision: applications may route work between real-time and more deliberative models rather than standardize on one endpoint.
OpenAI’s platform is moving toward a portfolio of specialized model capabilities, where access policy and workload fit can matter as much as a model’s headline capability.
The trend: AI platforms are evolving from single-model APIs into managed portfolios that segment reasoning, cost-sensitive, and real-time workloads.
OpenAI o1 has started rolling out in the API, with all the features that developers expect: like structured outputs, image inputs, and a 200k context window [image]
We're bringing OpenAI o1 to the API. We're rolling out access to developers on usage tier 5 starting today, and rollout will continue over the next few weeks. o1 supports: ⚙️ Function calling 🗂️ Structured Outputs 👀 Vision 📝 Developer messages 🧠 Reasoning effort [video]
Our reasoning model @OpenAI o1 is now in the API! It comes with function calling, developer messages, Structured Outputs, and vision. 🍓 We also shipped WebRTC support for the Realtime API, Preference Fine-Tuning, official Go/Java SDKs, and more! 🎁 https://openai.com/...
o1 being available via API may not be as sexy as video stuff but I predict that this will in retrospect be seen as the beginning of a significant increase in AI's impact on the economy. Many more use cases are now possible.
Sooooo new o1 API today from @OpenAI. Larger context window up to 200k and structured output. Feels like they finally caught up to @AnthropicAI. Is it time for me to switch to the new o1-2024-12-17 for coding? [image]
YAY!!! OpenAI comes through for Developers!! The o1-API is finally out, and it has fancy bells and whistles like function-calling 💃💃 AND JUST LIKE THAT - OpenAI overtakes Google again. NO experimental blah blah... The API IS GENERALLY AVAILABLE
OpenAI's o1 model is officially out of preview and available to developers via API! Plus, new capabilities for Realtime API builders (like the cool stuffed animal voice demo below). New o1 features include: > Function calling to connect with external data > Developer messages [vi…
I'ma be real. I don't want o1 in the API. o1 is very clearly an opaque chain of thought that summarises back into the context window. I want access to the model so I can use my own, better chain of thought system.
Little detail on o1 vs o1-pro I hadn't heard before, pro “uses techniques that go beyond thinking for longer” o1-pro api support coming (some time) [image]
new stuff for developers today, including o1 in the API with new features like structured outputs and function calling, realtime API updates, preference fine-tuning, and more.