ChatGPT version history runs from a research paper in 2018 to the GPT-6 family, which became the default model in ChatGPT on 7 October 2026.
More than twenty named models sit between those two points, and the gap between them keeps shrinking. Most version lists you find are a year out of date, which is a problem when OpenAI shipped four GPT-6 releases in five weeks.
We checked every date below against OpenAI’s release index and ChatGPT release notes in October 2026. This guide gives you every model, its release date, and what it changed, starting with the full table. Then it explains each generation and shows you how to tell which version you are on right now.
ChatGPT Version History at a Glance
Every ChatGPT and GPT model in order with its release date.
| Model | Released | What changed |
|---|---|---|
| GPT-1 | June 2018 | The first generative pre-trained transformer. A research paper and not a product. |
| GPT-2 | February 2019 | Bigger and far more coherent. OpenAI staged the release over misuse worries. |
| GPT-3 | June 2020 | 175 billion parameters. The first model that wrote usable long-form text. |
| GPT-3.5 | November 2022 | The model ChatGPT launched on, and how most people met AI. |
| GPT-4 | 14 March 2023 | Took images as well as text. Passed exams GPT-3.5 failed. |
| GPT-4o | 13 May 2024 | One model for text, vision and audio, with much faster replies. Left ChatGPT on 13 February 2026. |
| GPT-4o mini | 18 July 2024 | A smaller and cheaper GPT-4o. |
| o1 | 12 September 2024 | First reasoning model. It thought before answering. |
| o3-mini | Early 2025 | Cheaper reasoning, aimed at math and code. |
| GPT-4.5 | 27 February 2025 | Codenamed Orion. Better conversation with no reasoning jump. Left ChatGPT on 26 June 2026. |
| GPT-4.1 | April 2025 (API), 14 May 2025 (ChatGPT) | Built for developers and coding. Left ChatGPT on 13 February 2026. |
| o3 and o4-mini | 16 April 2025 | Reasoning models in ChatGPT. Last of the separate o-series line. |
| GPT-5 | 7 August 2025 | A router that picks a fast or a slow answer for you. |
| GPT-5.1 | 12 November 2025 | Warmer tone and better at following instructions. Left ChatGPT on 11 March 2026. |
| GPT-5.2 | 11 December 2025 | Stronger writing and data analysis. Left ChatGPT on 12 June 2026. |
| GPT-5.3 Instant | 3 March 2026 | More accurate answers, better web results and fewer dead ends. |
| GPT-5.4 Thinking | 5 March 2026 | Best coding scores of the GPT-5 line at the time. |
| GPT-5.5 | 23 April 2026 | Rolled out to the paid plans. Became the default Instant model on 5 May 2026. |
| GPT-5.6 | 9 July 2026 | Shipped in three sizes: Luna, Terra and Sol. Previewed on 26 June 2026. |
| GPT-6 Astra | 3 September 2026 | Top model. New architecture and a one million token context. |
| GPT-6 Sol and GPT-6 Luna | 22 September 2026 | Cheaper GPT-6 models for everyday and high-volume work. |
| GPT-6.1 Sol | 29 September 2026 | Near-Astra coding and computer use at one fifth of Astra’s API price. |
| GPT-6 in ChatGPT | 7 October 2026 | Default for ChatGPT users, with Intelligent UI. Sol for paid plans, Luna for Free and Go. |
Two things stand out in that list. The gap between big releases keeps shrinking, from years down to weeks. And since GPT-5, the point releases carry most of the real change. A version list that stops at GPT-5 tells you almost nothing about what you are using today.
The Early Models: GPT-1 to GPT-3
None of these were products. They were papers with model weights attached, and almost nobody outside research used them.
GPT-1 arrived in June 2018 and proved the idea: train a model on a lot of text and then adapt it to specific jobs. GPT-2 followed in February 2019 and wrote well enough that OpenAI held back the full version for months over fears of misuse. That decision looks quaint now and was taken seriously at the time.
GPT-3 landed in June 2020 with 175 billion parameters. It was the first one that could write a page of text you might not spot as machine-written. As AI-generated writing becomes increasingly difficult to distinguish from human writing, tools that help readers check if text is AI-generated are becoming more useful for evaluating content authenticity. Developers built demos on GPT-3 for two years before the public caught on.
GPT-3.5 and the Launch That Changed Everything
ChatGPT launched in November 2022 on GPT-3.5. The model was not much better than GPT-3. The chat window was the change.
A text box instead of an API key took the user count from a few thousand developers to a hundred million people. That took two months. Worth remembering when you read about model releases: the interface did more here than the weights did.
GPT-4 and the Move to Multimodal
GPT-4 shipped on 14 March 2023 and was the first real jump in ability since GPT-3. It accepted images, held far longer conversations, and passed professional exams that GPT-3.5 failed outright.
GPT-4o followed on 13 May 2024. The o stands for omni. One model now handled text, vision and audio together instead of stitching separate systems, and it replied fast enough for live voice conversation. For a lot of people GPT-4o is still the version they picture when they think of ChatGPT. OpenAI retired it from ChatGPT on 13 February 2026, together with GPT-4.1 and o4-mini (release notes).
GPT-4o mini arrived on 18 July 2024 as a cheaper sibling. GPT-4.1 reached the API in April 2025 and ChatGPT’s paid plans on 14 May 2025 (OpenAI announcement).
GPT-4.5, codenamed Orion, arrived on 27 February 2025. It was the best conversational model OpenAI had made and not a reasoning step forward. OpenAI retired it from ChatGPT on 26 June 2026, according to its release notes.
The o-Series: A Separate Line for Reasoning
For about seven months, OpenAI ran two model lines at once, and this is the part of the version history that confuses people most.
o1 arrived on 12 September 2024 and worked differently from every GPT model before it. Instead of predicting an answer straight through, it worked through steps first and then answered. It was slower and more expensive but much better at math and logic.
o3-mini and o4-mini followed through early 2025 as cheaper versions aimed at math and code. o3 and o4-mini reached ChatGPT on 16 April 2025. OpenAI retired o4-mini there on 13 February 2026 and set o3 to retire on 26 August 2026. Then the line ended. GPT-5 folded that step-by-step approach into the main model. There was no longer a reason to choose between a fast model and a thinking one.
The GPT-5 Line: 2025 to 2026
GPT-5 launched on 7 August 2025 and fixed the problem every regular user had: choosing which model to use. GPT-5 is a system of models with a router that reads your question and picks a quick answer or a longer thought.
The launch was rocky. People reported odd results and disliked losing GPT-4o, which OpenAI had retired the same day. Access to the older models came back for paid users within weeks.
After that came point releases every few weeks. GPT-5.1 on 12 November 2025 softened the tone. GPT-5.2 on 11 December 2025 improved writing and data work.
GPT-5.3 Instant on 3 March 2026 gave more accurate answers and fewer dead ends. GPT-5.4 Thinking on 5 March 2026 brought the best coding scores of the line. Dates come from OpenAI’s ChatGPT release notes.
GPT-5.5 arrived on 23 April 2026 for the paid plans (OpenAI announcement). It became the default Instant model on 5 May 2026.
GPT-5.6 followed on 9 July 2026, after a preview on 26 June. It came in three sizes named Luna, Terra and Sol, from cheapest to most capable (OpenAI announcement). That naming carried into the next generation.
Most of these models have already left ChatGPT. OpenAI says models generally stay available for 90 days after a successor ships (release notes). The list below shows what has been retired so far.
GPT-6 Astra: The Top GPT-6 Model
OpenAI released GPT-6 Astra on 3 September 2026, first to a limited set of organizations and then to paid users over the following days. It is available to ChatGPT Plus, Pro, Business and Enterprise users, through the OpenAI API as gpt-6-astra, on Microsoft Azure and on Amazon Bedrock.
What is new
- Context window: about 1,050,000 tokens, with up to 128,000 tokens of output. Long enough to hold a small codebase or a stack of contracts at once.
- Knowledge cutoff: 30 April 2026.
- Architecture: a design OpenAI calls recurrent depth, or looped transformers. It is more efficient, and OpenAI has said it hides some of the model’s reasoning steps.
- Training scale: pretrained on more than 100,000 GPUs at OpenAI’s Stargate site in Texas.
- Price: $10 per million input tokens and $50 per million output. Cached input drops to $1, batch runs at half price, and a Fast mode costs double.
How it scores
OpenAI reports 97.6% on FrontierMath Tier 4 and 96.0% on GPQA Diamond. On computer and browser tasks, it scores 72.6% on OSWorld 2.0, at roughly 47% less time per task than GPT-5.6 Sol.
Read those with care. The headline 99.9% on ARC-AGI-3 relies on a stateful and costly test setup, and plain API calls score much lower. Vendor benchmarks show what a model can do under ideal conditions. They do not show what you get by default.
The part that got the most attention
OpenAI is releasing GPT-6 Astra in stages, and its advanced cybersecurity abilities are held back behind an approved-tester program. The public version refuses certain security prompts. This is the first time OpenAI has gated a whole capability class at launch. It says more about where these models have got to than any score does.
GPT-6 Sol, Luna and GPT-6.1: The Latest ChatGPT Versions
The latest ChatGPT version is GPT-6, which began rolling out to every ChatGPT user on 7 October 2026. Plus, Pro, Business and Enterprise accounts get GPT-6 Sol. Free and Go accounts follow a day later with GPT-6 Luna.
The Pro reasoning option still uses GPT-6 Astra. OpenAI calls the new interface Intelligent UI, because answers can now come back as visuals and interactive tools (release notes).
- GPT-6 Sol and Luna (22 September 2026): Sol sits one tier below Astra and targets coding and complex work. Luna is the cheaper model for high-volume tasks. At launch they were in ChatGPT Work and Codex and the API. They were not in Chat. API prices per million tokens are $2 input and $10 output for Sol. Luna costs $0.10 input and $0.50 output (OpenAI announcement).
- GPT-6.1 Sol (29 September 2026): OpenAI says it reaches near-Astra results on coding and computer use. Its API price is one fifth of Astra’s standard rate: $2 input and $10 output per million tokens, as
gpt-6.1-sol(OpenAI announcement). - GPT-6 for everyone (7 October 2026): GPT-6 began rolling out as the default in ChatGPT worldwide (OpenAI announcement). It gives faster answers and can return diagrams, calculators and small tools you can use in the chat.
OpenAI’s API docs now point developers to GPT-6 Astra for complex reasoning and coding. They suggest GPT-6.1 Sol for balance and GPT-6 Luna for high-volume work (OpenAI model catalog). Check the OpenAI release index for anything newer.
Which ChatGPT Version Are You Using?
Most people cannot answer this, and it is the question that matters more than the history.
- Open the model picker: In the ChatGPT web app or mobile app, the model name sits at the top of the conversation. Tap it to see what is available on your plan.
- Check your plan and not the news: A launch announcement does not mean the model is on your account yet. OpenAI rolls new models out in stages, and free accounts get them last or not at all. If your plan sits a generation behind, a few free ChatGPT alternatives are worth comparing.
- Do not ask the model: Ask ChatGPT which version it is and it will often get it wrong. A model’s knowledge cutoff sits before its own release. It cannot know its own name.
- In the API, pin the version: Use the exact model ID (such as
gpt-6-astra) instead of an alias that moves. Behavior shifts between point releases and a moving alias will change your output with no warning.
That last one catches teams out constantly. A prompt tuned on GPT-5.3 can behave differently on GPT-5.4, and nothing in your code changed. Pin the version and test before you move. Each new version can change which sources ChatGPT cites. Track your AI search visibility after a move.
Conclusion
The ChatGPT version history is now more than twenty models in eight years, and the pace is still rising. GPT-1 to GPT-3 took two years each. GPT-5 to GPT-6 took thirteen months, and GPT-6 then shipped four releases in five weeks.
Two practical takeaways. Check which version your account is on before you judge ChatGPT. Free and paid accounts often sit a generation apart. And if you build on the API, pin your model version. A quiet update will otherwise change your results for you. OpenAI also retires old models. Check its release notes before you depend on one.
For how the current flagship stacks up against the leading open-weight model, see our comparison of DeepSeek vs ChatGPT.
Frequently Asked Questions (FAQs)
Q1. What is the latest ChatGPT version?
GPT-6, which became the default for every ChatGPT user on 7 October 2026. In ChatGPT it runs on GPT-6 Sol and GPT-6 Luna. GPT-6 Astra, released on 3 September 2026, is still the most capable model and is sold to paid plans and through the API.
Q2. How is GPT-6 Astra different from GPT-5.6?
It uses a different architecture that OpenAI calls recurrent depth, and its context window is about one million tokens. It also scores higher on reasoning and computer-use tests. It also costs more: $10 per million input tokens against $4 for GPT-5.6 Sol at its current price.
Q3. What happened to the o1, o3 and o4 models?
They were a separate reasoning line that ran alongside the GPT models from September 2024 to April 2025. GPT-5 absorbed their step-by-step approach into the main model. So the o-series was retired instead of continued.
Q4. Why does ChatGPT tell me the wrong version number?
Because a model’s training data ends before the model itself is released, it has no reliable information about its own name or release date. Check the model picker in the app instead of asking the model itself.
Q5. How many ChatGPT versions are there in total?
More than twenty named models between GPT-1 in June 2018 and GPT-6.1 Sol in September 2026, counting the o-series reasoning models and every point release.
Q6. Is GPT-4o still available?
No. OpenAI retired GPT-4o from ChatGPT on 13 February 2026, together with GPT-4.1, GPT-4.1 mini and o4-mini. Check the model picker on your own plan for what is available now.
Q7. Does a newer ChatGPT version always give better answers?
No. Newer models score higher on tests, but tone and formatting change between releases. A prompt tuned for an older model can perform worse on a newer one. Test your own prompts instead of assuming an upgrade helps.
