OpenAI for Business

We recently published a tactical, 30+ page playbook to help teams move from building LLM apps to deploying full-fledged agents. In "A practical guide to building agents", you’ll find: 🤖 Frameworks for deciding when (and when not) to use agents 🏅 Best practices for model selection, tool integration, and guardrails 🎨 Design patterns for single- and multi-agent systems 🖥️ Code examples using the OpenAI Agents SDK Whether you're focused on internal automation or customer-facing features, this guide will help you build agents that are safe, scalable, and production-ready. Check out the full guide below or at this link: https://lnkd.in/d_e2FP2u

Mercari, Inc. is one of Japan’s leading online marketplaces, helping users buy and sell secondhand goods. To make the listing process simpler and drive more sales, Mercari turned to OpenAI. By integrating models like GPT-4o mini into the listing process, sellers can now generate product listings just by uploading photos, making it faster and easier to start selling. This has led to: ⚡️ Hundreds of AI-assisted listings created every minute 📈 Higher conversion rates and increased per-user sales 🛍️ Easier, more accessible experiences—especially helpful for first-time sellers Mercari is also building on this foundation with reusable AI components, new customer-facing features, and internal productivity tools across the broader Mercari Group. 👇 Full story in the comments 👇

Today, we’re excited to announce improved citations in ChatGPT. ChatGPT can now include multiple citations for a given response, allowing you to learn more or verify information across more sources. We’ve also added a new ‘highlight’ UI to more clearly show which part of the answer the citation refers to. See how it works below.

OpenAI just launched image generation in the API, and we’re already testing it for marketers on HubSpot. What’s exciting is the practical power. OpenAI’s image generation capabilities are really good at rendering text accurately and transforming uploaded images. This means clear use cases for sales and marketing collateral that need your logo or brand assets. We're still in exploration mode, but staying on top of model updates is a big part of how we make AI easy. Had to try it on myself in the meantime…which one is best?

Starting today, Enterprise and EDU teams can experiment with OpenAI o3 and OpenAI o4-mini, our newest reasoning models in ChatGPT. We taught them to reason and act: Deciding when to use tools like search, Python, and charts to solve complex problems from start to finish. It’s a step toward a more agentic ChatGPT that can run tasks end-to-end on your behalf. Here’s a quick demo: ChatGPT tackles a budget variance report using search, code, and data visualization - all in under 2 minutes.

Image gen is now available in the API! We’re launching gpt-image-1, making ChatGPT’s powerful image generation capabilities available to developers worldwide starting today. ✅ More accurate, high fidelity images 🎨 Diverse visual styles ✏️ Precise image editing 🌎 Rich world knowledge 🔤 Consistent text rendering Check out the guide to start building: https://lnkd.in/gn-RGBji

Join us this Thursday at 11 AM for a 30-minute webinar to see image generation and our newest o3 model in action. We’re shipping new ChatGPT features every month: building a more agentic assistant tailored to how your teams work. We'll feature 5 demos on how your team can: 📁 Access and synthesize internal knowledge from Google Drive directly inside ChatGPT 🧠 Draft and collaborate naturally with GPT-4o 🔁 Automate complex workflows with our newest model, o3 🎯 Personalize ChatGPT interactions with improved memory 🎨 Create professional visuals for campaigns and decks Live attendees get prompt packs, how-to guides, and real-world examples to start using these features right away. 👉 Save your spot here: https://lnkd.in/gHAQu2dS

Speak CEO and co-founder Connor Zwick is building an advanced AI language tutor designed to help learners achieve fluency through natural, open-ended conversations. AI is core to how Speak helps users not just study a language—but actually learn to speak it confidently. Connor and his team think ahead, developing solutions that might seem technically challenging or expensive today but are likely to become feasible in the near future. 👇 Check out the full story in the comments below to see how Speak combines deep technical expertise with thoughtful product design to personalize language learning with AI. 👇

Yesterday, we announced OpenAI o3 and o4-mini - the latest in our o-series of models, available now in ChatGPT Team and coming soon to ChatGPT Enterprise. Check out the demo below to see an example of how a business could put o3 and o4-mini to work.

We’re excited to announce OpenAI o3 and o4-mini—the latest in our o-series of models trained to think for longer before responding, available today in ChatGPT Team and coming soon to ChatGPT Enterprise. These models offer advanced intelligence with capabilities like visual perception and analysis, and enhanced tools for search, data analysis, and canvas. They can combine tools seamlessly to take on more complex, longer-running tasks. They are also great at solving complex math, coding, and scientific challenges. 👇 Link to the full announcement in the comments below 👇