Table of Contents
Table of Contents
sources: OpenAI API Changelog / OpenAI News
2026
#| Date | Model/Product | Core features and breakthroughs |
|---|
| 2026-04-23 | GPT-5.5 / GPT-5.5 Pro | OpenAI introduced GPT-5.5 as its smartest and most intuitive model, rolling out in ChatGPT/Codex for Plus/Pro/Business/Enterprise tiers, with GPT-5.5 Pro for Pro/Business/Enterprise. Source: Introducing GPT-5.5 |
| 2026-04-21 | ChatGPT Images 2.0 | OpenAI released ChatGPT Images 2.0 and framed it as “a new era of image generation.” |
| 2026-03 (rollout) | GPT-5.4 Thinking (ChatGPT) | ChatGPT began rolling out GPT-5.4 Thinking (using the “Thinking” naming on ChatGPT); targeted at reasoning, coding, and agentic tasks. Note: ChatGPT naming and rollout cadence may not fully match API model names. |
| 2026-03 (rollout) | GPT-5.4 mini (ChatGPT) | Rolled out gradually in ChatGPT; used as a fallback in some plans/scenarios when GPT-5.4 Thinking hits rate limits. |
| 2026-03-17 | gpt-5.4-mini / gpt-5.4-nano | Released gpt-5.4-mini and gpt-5.4-nano (Chat Completions / Responses): mini targets high-concurrency, cost-efficient scenarios and supports tool search / computer use / compaction; nano targets simpler high-concurrency tasks, supports compaction only. |
| 2026-03-05 | gpt-5.4 / gpt-5.4-pro | Released gpt-5.4 (Chat Completions / Responses) and gpt-5.4-pro (Responses), adding tool search, built-in computer use, a (1M)-token context window, and native compaction for long conversations and agent workflows. |
| 2026-03-03 | gpt-5.3-chat-latest | Released gpt-5.3-chat-latest (Chat Completions / Responses API), mapped to the latest GPT‑5.3 (Instant) snapshot currently used in ChatGPT. |
| 2026-02-24 | gpt-5.3-codex | Released gpt-5.3-codex (Responses API). |
2025
#| Date | Model/Product | Core features and breakthroughs |
|---|
| 2025-12-11 | GPT-5.2 | Released GPT-5.2 (Instant/Thinking/Pro), and updated the knowledge cutoff date (officially noted as Aug 2025). |
| 2025-11-12 | GPT-5.1 | GPT-5 was updated to 5.1 Instant/Thinking, with stronger style control and personalization options. |
| 2025-08-07 | GPT-5 | New flagship model rolled out gradually; paid users could choose GPT-5 or GPT-5 Thinking, and Pro/Team had Thinking Pro. |
| 2025-06-10 | OpenAI o3-pro | Longer-thinking variant to improve reliability; released for Pro users in ChatGPT and also available via API. |
| 2025-05-14 | GPT-4.1 / GPT-4.1 mini | Coding-focused models entered ChatGPT; 4.1 mini replaced 4o mini. |
| 2025-04-16 | OpenAI o3 / o4-mini | Reasoning model update: o3 for stronger reasoning, and o4-mini for better cost-performance and higher usage limits. |
| 2025-02-27 | GPT-4.5 (research preview) | More natural conversational experience, stronger instruction adherence, and lower hallucination rate. |
| 2025-01-31 | o3-mini | Cost-efficient reasoning model optimized for coding, math, and science, with developer features such as structured outputs and tools. |
2024
#| Date | Model/Product | Core features and breakthroughs |
|---|
| 2024-09-12 | o1-preview / o1-mini | First release of the reasoning series, targeting more complex math/coding/science tasks. |
| 2024-07-18 | GPT-4o mini | Launched as a more cost-efficient small model for ChatGPT and developer scenarios. |
| 2024-05-13 | GPT-4o (Omni) | Native multimodal flagship covering text/vision/audio, with low-latency real-time interaction. |
| 2024-02-15 | Sora (first reveal) | Initial official reveal of the text-to-video model (capable of generating videos up to about 1 minute). |
2023
#| Date | Model/Product | Core features and breakthroughs |
|---|
| 2023-11-06 | GPT-4 Turbo | Released at DevDay with a larger context window and newer knowledge cutoff. |
| 2023-03-14 | GPT-4 | Released; first introduced image input support (with text output). |
2022
#| Date | Model/Product | Core features and breakthroughs |
|---|
| 2022-11-30 | ChatGPT (GPT-3.5) | Public research preview launch. |
There are no articles to list here yet.