A public timeline of major OpenAI GPT model releases, ChatGPT product updates, and related OpenAI milestones.
This repository is built as a clean reference for ChatGPT and GPT release history. It covers GPT-1, GPT-2, GPT-3, ChatGPT, GPT-4, GPT-4o, OpenAI reasoning models, GPT-5, Codex, Sora, ChatGPT Images, GPT-5.5, GPT-5.6, Computer History, ChatGPT for Teens, and GPT-5.6-Cyber.
Last updated: August 19, 2026
- GPT model release dates
- ChatGPT product milestones
- OpenAI reasoning model releases
- Codex and coding-agent updates
- Sora and video generation releases
- ChatGPT Images and image generation updates
- ChatGPT memory and personalization updates
- OpenAI Daybreak and cybersecurity model releases
- Major OpenAI API and developer platform milestones
| Model or Product | Release Date | Key Highlight |
|---|---|---|
| ChatGPT for Teens | August 18, 2026 | Learning-focused ChatGPT experience for teens with stronger default protections, parent controls, and healthy-use features. |
| GPT-5.6 Sol Ultrafast mode preview | August 14, 2026 | Preview mode that runs GPT-5.6 Sol at up to 14 times the speed. |
| ChatGPT Computer History | August 14, 2026 | Optional macOS desktop feature that turns activity across selected apps and websites into memories and a timeline for future interactions. |
| GPT-5.6-Cyber | August 10, 2026 | Cybersecurity-specific model for approved Daybreak Red users conducting advanced, authorized security work. |
| GPT-5.6 Sol, Terra, and Luna | June 27, 2026 | Limited preview of the GPT-5.6 series, with Sol as the flagship model, Terra for balanced work, and Luna for fast lower-cost work. |
| GPT-5.5 | April 23, 2026 | Frontier model for coding, research, computer use, documents, spreadsheets, and long-running work. |
| ChatGPT Images 2.0 | April 21, 2026 | New ChatGPT image generation experience and model generation. |
| GPT-5.4 mini and nano | March 17, 2026 | Smaller GPT-5.4 models for faster and lower-cost workloads. |
| GPT-5.4 | March 5, 2026 | Frontier model for professional work across reasoning, coding, agents, and tool use. |
| GPT-5.3-Codex-Spark | February 12, 2026 | Ultra-fast research preview model for real-time coding in Codex. |
| GPT-5.3-Codex | February 5, 2026 | Agentic coding model for professional software and computer-based work. |
| GPT-5.2-Codex | December 18, 2025 | Agentic coding model for complex software engineering and defensive cybersecurity. |
| GPT Image 1.5 | December 16, 2025 | Improved ChatGPT Images model with faster generation and stronger editing. |
| GPT-5.2 | December 11, 2025 | Frontier model series for professional knowledge work and long-running agents. |
| GPT-5.1 | November 12, 2025 | ChatGPT update with GPT-5.1 Instant and GPT-5.1 Thinking. |
| Sora 2 | September 30, 2025 | Video and audio generation model with stronger physics, realism, and control. |
| GPT-5 | August 7, 2025 | Unified GPT-5 system for ChatGPT and developers. |
| Codex | May 16, 2025 | Cloud-based software engineering agent. |
| OpenAI o3 and o4-mini | April 16, 2025 | Reasoning models with tool use across ChatGPT. |
| GPT-4.1 | April 14, 2025 | API model series for coding, instruction following, and long context. |
| 4o image generation | March 25, 2025 | Native image generation in GPT-4o and ChatGPT. |
| GPT-4.5 | February 27, 2025 | Research preview of a larger GPT model for chat. |
| OpenAI o3-mini | January 31, 2025 | Cost-efficient reasoning model for STEM, coding, and math. |
| OpenAI o1-preview | September 12, 2024 | First public reasoning model series from OpenAI. |
| ChatGPT search | October 31, 2024 | Web search experience inside ChatGPT. |
| Canvas | October 3, 2024 | Separate workspace for writing and coding with ChatGPT. |
| GPT-4o | May 13, 2024 | Omni model for text, audio, vision, and real-time interaction. |
| ChatGPT plugins | March 23, 2023 | Early tool system for ChatGPT, later deprecated. |
| GPT-4 | March 14, 2023 | Large multimodal model with text and image input. |
| ChatGPT Plus | February 1, 2023 | Paid ChatGPT subscription plan. |
| ChatGPT | November 30, 2022 | Public research preview of ChatGPT. |
| GPT-3 | May 28, 2020 | 175B-parameter language model for few-shot learning. |
| GPT-2 | February 14, 2019 | Large unsupervised language model and staged release milestone. |
| GPT-1 | June 11, 2018 | Early generative pre-training system for language understanding. |
| Date | Milestone |
|---|---|
| August 18, 2026 | OpenAI introduced ChatGPT for Teens, a learning-focused experience with Study Mode, responsible homework reminders, stronger age-appropriate protections, parent controls, and healthy-use features. |
| June 2018 | OpenAI published the first GPT research system through work on generative pre-training for language understanding. |
| February 2019 | GPT-2 was introduced as a larger unsupervised language model. OpenAI used a staged release process because of misuse concerns. |
| May 2020 | GPT-3 was introduced as a 175B-parameter language model that showed strong few-shot learning. |
| November 30, 2022 | ChatGPT launched as a public research preview. |
| February 1, 2023 | ChatGPT Plus launched as a paid subscription plan. |
| March 14, 2023 | GPT-4 launched as a multimodal model with text and image input. |
| March 23, 2023 | ChatGPT plugins launched as an early way for ChatGPT to use tools and external services. |
| July 2023 | GPT-4 API access became generally available to paying API customers. |
| May 13, 2024 | GPT-4o launched as a faster omni model for text, audio, vision, and real-time interaction. |
| September 12, 2024 | OpenAI introduced o1-preview, the first model in its reasoning series. |
| October 3, 2024 | Canvas launched as a separate workspace for writing and coding with ChatGPT. |
| October 31, 2024 | ChatGPT search launched for web answers with source links. |
| January 31, 2025 | OpenAI o3-mini launched as a smaller reasoning model for STEM, math, and coding. |
| February 27, 2025 | GPT-4.5 launched as a research preview of a larger GPT model for chat. |
| March 25, 2025 | OpenAI introduced native image generation in GPT-4o. |
| April 14, 2025 | GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano launched in the API. |
| April 16, 2025 | OpenAI o3 and o4-mini launched as reasoning models with access to ChatGPT tools. |
| May 16, 2025 | Codex launched as OpenAI's cloud-based software engineering agent. |
| August 7, 2025 | GPT-5 launched as a unified system in ChatGPT and the API. |
| September 30, 2025 | Sora 2 launched as a video and audio generation model. |
| November 12, 2025 | GPT-5.1 launched in ChatGPT as GPT-5.1 Instant and GPT-5.1 Thinking. |
| November 13, 2025 | GPT-5.1 became available for developers in the API. |
| December 11, 2025 | GPT-5.2 launched for professional knowledge work and long-running agents. |
| December 16, 2025 | GPT Image 1.5 launched through the new ChatGPT Images experience. |
| December 18, 2025 | GPT-5.2-Codex launched for agentic coding and cybersecurity-related software work. |
| February 5, 2026 | GPT-5.3-Codex launched for long-running coding, research, and computer-based work. |
| February 12, 2026 | GPT-5.3-Codex-Spark launched as an ultra-fast research preview for real-time coding. |
| March 5, 2026 | GPT-5.4 launched in ChatGPT, the API, and Codex for professional work. |
| March 17, 2026 | GPT-5.4 mini and GPT-5.4 nano launched for faster and lower-cost use cases. |
| April 21, 2026 | ChatGPT Images 2.0 launched as a new image generation experience. |
| April 23, 2026 | GPT-5.5 launched for advanced coding, research, computer use, knowledge work, and agentic workflows. |
| April 24, 2026 | GPT-5.5 and GPT-5.5 Pro became available in the API. |
| June 27, 2026 | OpenAI began a limited preview of the GPT-5.6 series: Sol, its flagship model; Terra, a balanced model for everyday work; and Luna, a fast and affordable model. Initial access is through the API and Codex for selected trusted partners. |
| August 10, 2026 | OpenAI introduced GPT-5.6-Cyber through Daybreak Red for approved defenders conducting advanced vulnerability research, exploit validation, and authorized security testing. The expanded Daybreak program also added Daybreak Blue access to GPT-5.6 Sol for defensive security work. |
| August 14, 2026 | ChatGPT introduced Computer History in the macOS desktop app. The optional feature turns activity across selected apps and websites into memories and a timeline that ChatGPT and Codex can reference later. |
| August 14, 2026 | OpenAI previewed Ultrafast mode for GPT-5.6 Sol at up to 14 times the speed. |
ChatGPT for Teens is a learning-focused experience for teens, with Study Mode, responsible homework reminders, Quizzes, Learning Visualizations, and Study Hours. OpenAI says its system automatically places users it estimates are under 18, as well as users who state that they are between 13 and 17, into the experience.
The experience applies stronger age-appropriate protections by default, including safeguards for self-harm, violence, eating disorders, dangerous activities, and explicit sexual or graphic content. Linked parents can set Quiet Hours, manage selected settings, and receive limited safety notifications. Break reminders, sensitive-image upload reminders, teen-specific onboarding, and clear cues that identify ChatGPT as AI support healthier use.
Computer History turns activity across selected apps and websites into memories and a timeline that ChatGPT and Codex can use in future interactions. This can help users resume recent work with less repeated explanation.
The feature is off by default in the ChatGPT desktop app on macOS. Pro users can choose to enable it. Business and Enterprise users need an administrator to grant workspace access before they can opt in. Computer History requires Memories and is not currently available in the European Economic Area, Switzerland, or the United Kingdom.
Users choose which apps and websites can contribute. Computer History records interaction events and creates text summaries, but it does not capture screenshots, screen recordings, microphone input, or system audio. Users can pause collection and inspect or delete their history.
OpenAI also previewed Ultrafast mode for GPT-5.6 Sol on August 14, 2026. The mode can run GPT-5.6 Sol at up to 14 times the speed.
The newest model release in this timeline is GPT-5.6-Cyber, introduced on August 10, 2026. It is a cybersecurity-specific model built on GPT-5.6 Sol and available through the restricted Daybreak Red program. It is intended for approved defenders conducting advanced vulnerability research, exploit validation, and authorized security testing.
OpenAI divided the expanded Daybreak program into two access tiers. Daybreak Blue provides approved defenders with general-purpose frontier models, including GPT-5.6 Sol, for defensive work such as secure code review, malware analysis, incident response, vulnerability discovery, and patch validation. Daybreak Red provides purpose-trained cyber models for higher-risk authorized work.
OpenAI reported that GPT-5.6-Cyber completed 95.0% of requests in its internal Advanced Cybersecurity Completion Rate evaluation, compared with 57.3% for GPT-5.5-Cyber. OpenAI assessed GPT-5.6-Cyber as High for cybersecurity capability but below its Critical threshold. Access requires approval and includes identity verification, account security, monitoring, approved-use restrictions, and legal attestations.
The latest major general-purpose GPT model in this timeline is GPT-5.6 Sol, introduced in limited preview on June 27, 2026.
GPT-5.6 Sol is OpenAI's strongest model in the announcement. The GPT-5.6 series also includes Terra for balanced everyday work, with competitive performance to GPT-5.5 at half the cost, and Luna as the fast, lowest-cost model in the series. During the preview, GPT-5.6 models are available through the API and Codex to selected trusted partners, with broader ChatGPT, Codex, and API availability planned soon. OpenAI says the short-term limited preview follows engagement with the U.S. government.
GPT-5.6 adds max reasoning effort for Sol and an ultra mode that can use subagents for complex work. OpenAI cites improved agentic capabilities in coding, biology, and cybersecurity, including Terminal-Bench 2.1, GeneBench v1, ExploitBench, and ExploitGym results. OpenAI says Sol did not cross the Cyber Critical threshold in the reported tests and that it used more than 700,000 A100-equivalent GPU hours for automated red teaming.
API pricing is listed per 1 million tokens: Sol at $5 input and $30 output, Terra at $2.50 input and $15 output, and Luna at $1 input and $6 output. GPT-5.6 also adds explicit cache breakpoints, a 30-minute minimum cache life, cache writes billed at 1.25x the uncached input rate, and a 90% cached-input discount for cache reads. OpenAI also plans to launch GPT-5.6 Sol on Cerebras at up to 750 tokens per second in July for selected customers.
This repository is maintained as a companion resource for the full article:
ChatGPT Timeline: GPT Release Dates From GPT-1 to GPT-5.6
The article includes a readable web version of the timeline, more context on GPT model generations, and related AI resources.
This repository is an independent timeline resource. It is not affiliated with OpenAI.
This repository is available under the MIT License.