Breaking: OpenAI Announces Revolutionary New Tools for Developers — A New Era for AI Agent Development
On March 12th (JST), OpenAI announced two new developer-focused features: the Responses API and the Agents SDK. This announcement marks a significant turning point in AI agent development. This article explains both tools in detail.
Topics covered:
- Background: Why These Tools Were Needed
- The New API Primitive: Responses API
- The Open-Source SDK: Agents SDK
- Summary
Background: Why These Tools Were Needed
OpenAI has already shipped two agents inside ChatGPT this year. The first is WebGPT, which browses the web and executes tasks on the user's behalf. The second is DeepResearch, which produces detailed reports on topics specified by the user. Feedback on both agents has been highly positive — and OpenAI's next step was to make these capabilities available to developers through an API.
Over the past several months, OpenAI collaborated with developers around the world to understand what would make agent development easier. The conversations surfaced a consistent theme: while the models themselves are ready to handle complex, multi-step workflows, developers face a painful process of stitching together various low-level APIs from different sources — a process that is difficult, time-consuming, and often brittle.
The Responses API and Agents SDK were built specifically to solve these problems — making agent development significantly more approachable.
Struggling with AI adoption?
We have prepared materials covering ZEROCK case studies and implementation methods.
The New API Primitive: Responses API
The Responses API is a new API primitive for building agents that leverage OpenAI's built-in tools. It supports three newly introduced integrated tools: Web Search, File Search, and Computer Use.
1. Web Search Tool Enables models to access up-to-date information from the internet, keeping responses current. The web search capability in the API uses the same model powering ChatGPT Search. On SimpleQA — a benchmark evaluating LLM accuracy on short factual questions — GPT-5.2 Search Preview and GPT-5.2 Mini Search Preview scored 90% and 88%, respectively.
2. File Search Tool Allows developers to upload and embed documents, making relevant information easily retrievable. Practical applications include:
- Customer support agents that can quickly access FAQ content
- Legal assistants that can rapidly reference past cases before escalating to specialists
- Coding agents that can query technical documentation
3. Computer Use Tool Enables control of computers with graphical user interfaces — including virtual machines and legacy applications — without requiring API access. Use cases include automating browser-based workflows such as web app QA testing and data entry tasks across legacy systems.
The Open-Source SDK: Agents SDK
OpenAI also announced the open-source Agents SDK — an evolved, production-ready version of the earlier experimental "Swarm" release, designed to make agent orchestration significantly easier.
The SDK includes:
- Agent functionality: Agents that are easy to configure with clear instructions and built-in tools
- Handoff capability: Intelligent transfer of control between agents when appropriate
With the Agents SDK, complex multi-agent applications — such as a customer support agent that delegates refund processing to one agent and billing inquiries to another — can be built cleanly and efficiently. The SDK is well-suited for a wide range of real-world applications including customer support automation, multi-step research, content generation, code review, and sales prospecting.
Summary
OpenAI plans to continue supporting the Chat Completions API while delivering new capabilities through the Responses API. The team will also be integrating Assistants API functionality into the Responses API and providing a migration guide to ensure a smooth transition for developers.
2026 is shaping up to be the year of agents. ChatGPT and OpenAI's developer tools are evolving from simply answering questions to executing real-world tasks on behalf of users. OpenAI will continue building powerful new models and the tools developers need to build even better agents. The new era of AI agents is here.
References:
- https://openai.com/index/new-tools-for-building-agents/
- https://www.youtube.com/watch?v=hciNKcLwSes&t=1s
ChatGPT Now Available by Phone and on WhatsApp
On December 19th (JST), OpenAI announced that ChatGPT is now accessible by phone and through WhatsApp — previously only available via web and mobile app. As long as you have a phone line, you can now have a spoken conversation with ChatGPT. It raises an interesting question: could AI eventually handle call center and phone operator duties as well?
ChatGPT by Phone — Voice Interaction Is Here
OpenAI has launched a service allowing users in the United States to access ChatGPT by calling 1-800-CHATGPT (1-800-242-8478). Users can have voice conversations with ChatGPT directly over the phone.
The demo showed examples such as getting Spanish greetings on behalf of a friend's grandmother who only speaks Spanish, and identifying a building just from a verbal description. ChatGPT handles a wide range of questions accurately, even in a phone call format.
This feature works over a standard phone line — meaning it's accessible even where data connectivity is unreliable. It can be used from mobile phones as well as landlines, including rotary phones.
OpenAI offers 15 minutes of free calling per month. Users who want longer conversations can create an account in the app to access extended usage. (Currently limited to US callers.)
ChatGPT on WhatsApp — Integration with a Global Messaging Platform
OpenAI has also integrated ChatGPT with WhatsApp, one of the most widely used messaging apps in the world. Users can add ChatGPT as a contact by entering the phone number 1-800-242-8478 in WhatsApp and then start messaging directly.
The demo showed a user asking ChatGPT for a pasta sauce recipe and then requesting vegan and meat-based variations.
The integration makes ChatGPT more accessible to a much larger global audience. Currently, the service supports text-based conversations only, but future updates via account authentication are expected to add features like image search.
OpenAI's Mission — AI for All of Humanity
OpenAI's Head of Product Kevin Weil described the company's mission:
"OpenAI's mission is to ensure that artificial general intelligence benefits all of humanity — and for that to happen, it's essential that as many people as possible have access to it."
Making ChatGPT available by phone and on WhatsApp is an important step toward realizing that mission. By offering a free version of ChatGPT that anyone can use — not just power users running advanced models — OpenAI is broadening access to the benefits of AI.
Summary
OpenAI's move to bring ChatGPT to phone and WhatsApp is an important step toward making AI more accessible. Voice calls and integration with a global messaging platform mean more people than ever can now use ChatGPT. It also opens the door to a future where AI handles phone-based customer service roles currently handled by human operators.
OpenAI's technology, including ChatGPT, is expected to continue enriching our daily lives in new ways.
Reference: Open AI official website "Day 10: 1-800-CHATGPT"
Chat with ChatGPT on Video — OpenAI Adds Video Call and Screen Share to ChatGPT
On December 12th (local time), OpenAI announced the addition of visual features (video call and screen share) and a Santa Mode to ChatGPT's "Advanced Voice" mode. These new capabilities allow AI to analyze visual information from a camera in real time — understanding context and detail that voice alone can't capture. This article explains what these new ChatGPT features bring to the table.
Video Call — More Natural Communication
OpenAI has added a video call feature to ChatGPT, enabling users to have face-to-face conversations with the AI in real time. In the demo, ChatGPT recognized a coffee maker and the user's movements through the video feed, providing detailed guidance on hand-drip coffee preparation — from filter setup to blooming to pouring technique — while answering questions accurately. In another demo, ChatGPT participated in group video introductions and instantly recalled which colleague was wearing a Santa hat.
This new feature goes beyond text-based communication, providing a more intuitive and hands-on learning experience. With facial expressions and physical gestures conveying additional information, conversations become richer and more natural.
To use the video call feature:
- Click the "Advanced Voice Mode" button in the lower right of the ChatGPT chat screen
- Click the "Video" button to start the video call
- Enjoy real-time conversation with ChatGPT
This feature uses OpenAI's native multimodal model — ingesting audio directly and outputting audio directly. As a result, conversation is possible in 50+ languages with emotional and tonal depth, at a natural conversational pace — enabling cross-language, global communication.
Screen Share — Clearer, More Context-Specific Explanations
The second new feature is screen sharing. Users can now share not just their phone camera feed, but also whatever is on their screen, with ChatGPT — enabling more specific and understandable explanations. In the demo, a user shared a messaging app screen, and ChatGPT recognized the app and messages on screen and suggested appropriate responses.
Screen share works the same way as the video call: click "Advanced Voice Mode," then select "Share Screen." Using screen share alongside ChatGPT makes communication smoother and problem-solving more efficient.
Video and screen share began rolling out on the 13th in the latest mobile app, with full availability to all Team plans and most Plus and Pro plan users over the following week. Enterprise plan users are expected to gain access in early the following year.
Limited-Time: Chat with Santa Claus
OpenAI offered a limited-time feature in December 2024: ChatGPT users could chat with Santa Claus. By clicking the snowflake icon displayed on the ChatGPT home screen, users could enjoy a conversation with Santa.
Topics available with Santa included:
- Questions about life at the North Pole
- What the elves are up to
- Fun Christmas stories
Santa responded in real time with the jolly, festive tone you'd expect — a delightful feature for the holiday season.
The feature became available worldwide on the 13th wherever ChatGPT's voice mode is supported. First-time users of the Santa feature also had their Advanced Voice usage limits reset — meaning users who had exhausted their daily or monthly limit could immediately talk to Santa in Advanced Voice mode. After the limit was reached again, users could continue chatting with Santa in standard voice mode.
Summary
OpenAI added video call and screen share to ChatGPT, creating richer, more interactive communication experiences. These new features enable cross-language global conversations and more efficient problem-solving. The limited-time Santa chat feature brought some holiday cheer along the way. OpenAI will continue expanding ChatGPT's capabilities and delivering better experiences for users.
Reference: Open AI official website "Day 6: Advanced voice with video & Santa mode"
ChatGPT Mobile App Hits $4.58M in Revenue — But Growth Is Slowing
Summary
ChatGPT's mobile app set a new record in September with 15.6 million downloads and $4.6 million in revenue. However, revenue growth rates have been gradually declining — falling to 20% in September. The $19.99/month ChatGPT+ subscription accounts for the majority of revenue but shows signs of market saturation. Competitor Ask AI spends heavily on advertising and still out-earns ChatGPT on mobile. The US remains ChatGPT's largest market, accounting for 60% of revenue.
Sarah Perez @sarahintampa / October 10, 2023
There is both good news and bad news about ChatGPT's mobile performance. On the positive side, the AI app continued its growth trajectory, setting new records in September with 15.6 million downloads across iOS and Android globally, and total revenue of approximately $4.6 million. However, new data from market intelligence firm Appfigures shows that revenue growth is beginning to slow. While monthly revenue growth exceeded 30% for several months — 31% in July and 39% in August — it fell to just 20% in September.
The slowdown may be an early signal that ChatGPT is approaching saturation among mobile users willing to upgrade to the paid ChatGPT+ subscription — a $19.99/month in-app purchase that offers faster response times, priority access during peak hours, and early access to new features. ChatGPT's mobile revenue totals to date: $2.1 million in June, $2.74 million in July, $3.81 million in August, and the record $4.58 million in September.
Interestingly, ChatGPT is not actually the top-earning AI app. Competitor Ask AI, which spends heavily on advertising, peaked at $6.55 million in August (up from $6.48 million in May when ChatGPT mobile launched) and, according to Appfigures data, dipped slightly to $5.51 million in September — still higher than ChatGPT. Other competitors like Genie and AI Chat Smith have not grown as significantly as Ask AI.
Of course, when factoring in Ask AI's advertising spend, net revenue tells a different story. For ChatGPT, after Apple and Google take their cuts of in-app purchase revenue, net earnings in September were approximately $3.2 million.
In addition to revenue records, ChatGPT reached 15.6 million installs in September, bringing its estimated lifetime total to 52.2 million according to Appfigures. Google Play drove the majority of downloads with 9 million in September, with the App Store accounting for the remaining 6.6 million.
As expected, the App Store drives the lion's share of revenue — $3 million of last month's total in-app purchases. The US is the largest market for AI apps, accounting for 60% of ChatGPT's revenue.
This AI news is produced by TIMEWELL.
TIMEWELL's AI Consulting
TIMEWELL is a professional team supporting business transformation in the age of AI agents.
Our Services
- AI Agent Implementation: Business automation using GPT-5.2, Claude Opus 4.5, and Gemini 3
- GEO Strategy Consulting: Content marketing strategy for the AI search era
- DX Promotion & New Business Development: Business model transformation through AI
In 2026, AI is evolving from a tool you use to a colleague you work with. Let's build your AI strategy together.
