Premium museum expansion · Provider catalog
OpenAI
Frontier GPT and o-series models behind ChatGPT, the API, and Codex-class coding agents. This page summarizes public product lines and release narratives (March 2026 snapshot); see OpenAI, GPT-4, and ChatGPT for encyclopedic detail.
Architecture & training (overview)
Consumer models are typically decoder-only transformers trained on broad internet, licensed data, and human feedback — with later generations adding multimodal inputs (vision, audio), tool use, and reasoning-time compute for the o-series (chain-of-thought style inference, not one giant frozen CoT in the weights). Exact parameter counts for GPT-4 and GPT-5-class models are often undisclosed; treat tables as directional.
Official hub: openai.com · API docs: Models.
Tiers & naming
- GPT-3.5 — fast, economical chat class; many SKUs retired over time.
- GPT-4 — flagship multimodal era (vision in GPT-4 Turbo / 4o paths).
- GPT-4o (“omni”) — unified multimodal model family (text/image/audio) at lower latency than prior GPT-4 variants in OpenAI messaging.
- o-series (o1, o3, …) — reasoning-optimized; emphasizes test-time compute and structured problem solving.
- GPT-5.x — successor generation in product naming (exact versioning follows OpenAI’s marketing and API IDs).
Launch & benchmarks (selected)
GPT-4 (March 2023) set a new bar for general capability vs GPT-3.5; multimodal and larger-context variants followed through 2023–2024. GPT-4o (May 2024) emphasized omnimodal interaction and latency. o1 (2024) introduced explicit reasoning-time scaling for math and code benchmarks.
Competitive framing in industry coverage often references Anthropic Claude, Google Gemini, and xAI Grok in the same cohort by era.
Model versions
ChatGPT consumer lineup and API / completion lineage are different retirement schedules — both tables below are chronological (oldest → newest). Verify live access in API models and ChatGPT settings.
ChatGPT lineup (product tiers)
Museum snapshot aligned with public ChatGPT model version tables (release dates and “status in ChatGPT” churn — not API availability). Red = discontinued in product, amber = legacy support, green = active default path.
| Version | Release date | Status in ChatGPT | Notes |
|---|---|---|---|
| GPT-3.5 | Nov 2022 | Discontinued | Original fast chat tier before GPT-4-class defaults. |
| GPT-4 | Mar 2023 | Discontinued | First GPT-4 in ChatGPT; superseded by 4o / later lines. |
| GPT-4o | May 2024 | Discontinued | Omni default era; cleared as GPT-5.x becomes frontline. |
| GPT-4o mini | Jul 2024 | Discontinued | Cost-focused 4o; rotated out with GPT-5 stack. |
| o1-preview / o1-mini | Sep 2024 | Discontinued | Early reasoning previews; replaced by later o / GPT-5 routing. |
| o1 / o1-pro | Dec 2024 | Discontinued | Full o1 + pro tier; succeeded by o3 / GPT-5 reasoning paths. |
| o3-mini / o3-mini-high | Jan 2025 | Discontinued | Fast o3-class; folded into newer menus. |
| GPT-4.5 | Feb 2025 | Discontinued | Large preview / bridge model; retired from picker. |
| GPT-4.1 / GPT-4.1 mini | Apr 2025 | Discontinued | 4.x refresh; superseded by GPT-5 generation. |
| o3 | Apr 2025 | Legacy support | Reasoning tier still reachable for some accounts while GPT-5.x rolls out. |
| o4-mini / o4-mini-high | Apr 2025 | Discontinued | Fast reasoning mini; cleared in favor of GPT-5 mini / nano. |
| o3-pro | Jun 2025 | Discontinued | High-budget reasoning; replaced in default paths. |
| GPT-5 | 7 Aug 2025 | Discontinued | Unified GPT-5 flagship; consumer default moved to 5.2+ wave. |
| GPT-5 mini | 7 Aug 2025 | Legacy support | Faster, cheaper GPT-5-class; still offered while 5.2+ is primary. |
| GPT-5.1 | 12 Nov 2025 | Legacy support | Personality / style options; prior to 5.2 default. |
| GPT-5.2 | 11 Dec 2025 | Active | Frontier consumer tier before 5.3/5.4 wave. |
| GPT-5.3-Codex | 5 Feb 2026 | Active | Codex-integrated coding / agent experiences in product. |
| GPT-5.3-Instant | 3 Mar 2026 | Active | Low-latency instant path alongside frontier tiers. |
| GPT-5.4 | 5 Mar 2026 | Active | Current professional / computer-use focused wave (March 2026 snapshot). |
API & completion line (chronological)
Developer-facing SKUs and eras — many models stay on the API after ChatGPT rotation. Ordered by first public generation.
| Name | Era (approx.) | API status | Notes |
|---|---|---|---|
| Codex (code completion) | Aug 2021 | Legacy | Early code-specialized API model; superseded by later GPT stacks. |
| InstructGPT / RLHF path | Jan 2022 | Superseded | Instruction-tuned GPT-3 class; precursor to ChatGPT / Turbo lineage. |
| GPT-3.5 Turbo | Mar 2022 → ongoing | GA / legacy | Fast chat class; many dated snapshots retired over time. |
| GPT-4 (base) | Mar 2023 | Superseded | First GPT-4 API generation; multimodal follow-ups in same era. |
| GPT-4 Turbo | Nov 2023 | Superseded | 128K context; vision-capable path. |
| GPT-4o | May 2024 | GA | “Omni” multimodal (text / image / audio) at lower latency than prior GPT-4 variants. |
| GPT-4o mini | Jul 2024 | GA | Cost- and latency-optimized 4o-class. |
| o1-preview / o1 | Sep–Dec 2024 | GA | First broad o-series; explicit reasoning-time compute in product UX. |
| o1-mini | Sep 2024 | GA | Smaller reasoning tier. |
| o3 | Dec 2024 → 2025 | GA | Next o-series reasoning generation; STEM / agent use cases. |
| o3-mini | Early 2025 | GA | Faster, cheaper o3-class. |
| GPT-4.5 (Preview) | Feb 2025 | Deprecated | Large preview; retired from menus in favor of later lines. |
| GPT-4.1 / GPT-4.1 mini / nano | Apr 2025 | GA | GPT-4-class refresh in API naming; smaller tiers for cost. |
| o4-mini | Apr 2025 | GA | Fast reasoning tier; some copy positions GPT-5 mini as successor in workflows. |
| o3-pro | 2025 | Prior | Higher reasoning budget o3 variant where offered. |
| GPT-5 | Aug 2025 | GA | Unified GPT-5 generation flagship; configurable reasoning effort. |
| GPT-5 mini / nano | Aug 2025 | GA | Cost- and latency-optimized GPT-5-class tiers. |
| GPT-OSS (gpt-oss-20b / 120b) | Aug 2025 | Open | Open-weight line for local and partner inference. |
| GPT-5.1 / GPT-5 pro | Late 2025 | GA | Iterative GPT-5.x; pro / chat-tuned variants in menus. |
| GPT-5.1 Codex / Codex mini / Codex Max | Late 2025 | Prior | Agentic Codex tiers before GPT-5.2-Codex in menus. |
| GPT-5.2 / GPT-5.2 pro | Dec 2025 | GA | Frontier + pro before 5.3/5.4 wave. |
| GPT-5.2-Codex | Dec 2025 → Feb 2026 | GA | Agentic coding line; long context in tool menus. |
| GPT-5.3-Codex | Feb 2026 | GA | Successor Codex-class in agentic coding pickers. |
| GPT-5.4 / GPT-5.4 pro / mini / nano | Mar 2026 | GA / rolling | Current frontier wave in March 2026 snapshot; sub-tiers in IDE/API lists. |
Technical notes (where public)
| Topic | Public detail |
|---|---|
| Parameters (GPT-4) | Not officially disclosed; rumors span wide ranges — omit from authoritative claims. |
| Context (GPT-4 Turbo class) | 128K tokens (product era; check current API docs for exact model IDs). |
| Modalities | GPT-4o and later omnimodal paths: text, image, audio in unified stacks. |
Selected timeline (2024–2026)
- 2024: GPT-4o; o1 reasoning line; broad ChatGPT UI refresh.
- 2025: GPT-5 launch; GPT-4.1; continued o-series; open-weight GPT-OSS narrative.
- 2026: GPT-5.x point releases; tight integration with coding agents and partner IDEs.
OpenAI wall of models (March 2026 snapshot)
Condensed from public ChatGPT / API model picker groupings (frontier, multimodal, Codex, research). Labels and availability change — verify in API model docs.
ChatGPT lineup (high level)
Active frontline SKUs commonly described as GPT-5.2, GPT-5.3-Codex, and GPT-5.4, with GPT-5 and GPT-5 mini in legacy support. Earlier GPT-4.1 / o3-pro / o4-mini tiers roll into legacy or sunset paths as GPT-5.x and refined o-series ship.
| Model | Status | Primary focus |
|---|---|---|
| GPT-5.4 | Active | Professional work, computer use (Mar 5, 2026 narrative). |
| GPT-5.3-Codex | Active | Software development; Codex (Feb 5, 2026 narrative). |
| GPT-5.2 | Active | General-purpose flagship (Dec 11, 2025 narrative). |
| GPT-5.1 | Legacy | Alternative personalities / prior GPT-5.1-era UX. |
| GPT-5 / GPT-5 mini | Legacy | General purpose; fast / cost-efficient (Aug 7, 2025 GPT-5 narrative). |
| o3 / o3-pro | Legacy | Advanced reasoning; superseded in menus by newer GPT-5.x + o-series. |
| GPT-4.1 series | Discontinued / legacy | Code and analysis era; cleared as GPT-5.x becomes default. |
Frontier models
“OpenAI’s most advanced models” bucket in product docs.
| Model | Notes |
|---|---|
| GPT-5.4 | Agentic, coding, professional workflows at scale. |
| GPT-5.4 pro | Smarter, more precise variant of GPT-5.4. |
| GPT-5.4 mini | Strong mini for coding, computer use, subagents. |
| GPT-5.4 nano | Cheapest GPT-5.4-class tier for high-volume simple tasks. |
| GPT-5 mini | Near-frontier; cost-sensitive, low-latency, high volume. |
| GPT-5 nano | Fastest, most cost-efficient GPT-5-class. |
| GPT-5 | Prior reasoning flagship; configurable reasoning effort. |
| GPT-4.1 | Smartest non-reasoning GPT-4.x-class in this menu. |
Image
| Model | Status | Notes |
|---|---|---|
| GPT Image 1.5 | Current | State-of-the-art image generation. |
| chatgpt-image-latest | Current | Image model used in ChatGPT. |
| GPT Image 1 | Current | Previous-generation image stack. |
| gpt-image-1-mini | Current | Cost-efficient GPT Image 1. |
| DALL·E 3 | Deprecated | Prior generation image model. |
| DALL·E 2 | Deprecated | First DALL·E generation. |
Video
| Model | Status | Notes |
|---|---|---|
| Sora 2 | Deprecated | Flagship video with synced audio (menu label). |
| Sora 2 Pro | Deprecated | Most advanced synced-audio video (menu label). |
Realtime & audio
| Model | Notes |
|---|---|
| gpt-realtime-1.5 | Voice: audio in / audio out (realtime). |
| gpt-realtime | Realtime text + audio I/O. |
| gpt-realtime-mini | Cost-efficient GPT Realtime. |
| gpt-audio | Audio I/O via Chat Completions API. |
| gpt-audio-1.5 | Best voice model for Chat Completions audio path. |
| gpt-audio-mini | Cost-efficient GPT Audio. |
| GPT-4o Audio | GPT-4o with audio inputs and outputs. |
| GPT-4o mini Audio | Smaller audio-capable 4o. |
| GPT-4o Realtime / GPT-4o mini Realtime | Realtime text + audio. |
| GPT-4o Transcribe / GPT-4o mini Transcribe | Speech-to-text (4o / 4o mini). |
| GPT-4o Transcribe Diarize | Transcription with speaker diarization. |
| GPT-4o mini TTS | Text-to-speech via 4o mini. |
| TTS-1 / TTS-1 HD | Fast vs quality text-to-speech. |
| Whisper | General speech recognition. |
Coding
| Model | Status | Notes |
|---|---|---|
| GPT-5.3-Codex | Current | Most capable agentic Codex model (menu copy). |
| GPT-5.2-Codex | Current | Long-horizon agentic coding. |
| GPT-5-Codex | Current | GPT-5 optimized for Codex. |
| GPT-5.1 Codex / GPT-5.1-Codex-Max / GPT-5.1 Codex mini | Prior tiers | Agentic Codex variants; mini is cost-focused. |
| codex-mini-latest | Deprecated | Fast Codex CLI-oriented model (superseded in menus). |
Deep research
| Model | Notes |
|---|---|
| o3-deep-research | Most powerful deep-research SKU. |
| o4-mini-deep-research | Faster, more affordable deep research. |
Open-weight models
| Model | Notes |
|---|---|
| gpt-oss-120b | Largest open-weight; H100-class single-GPU framing in docs. |
| gpt-oss-20b | Medium open-weight; low-latency use. |
More models (mixed API SKUs)
| Model | Status | Notes |
|---|---|---|
| GPT-5.2 / GPT-5.2 pro | Prior | Earlier GPT-5.2-era frontier + pro. |
| GPT-5.1 / GPT-5 pro | Prior | Coding / agentic GPT-5 generation. |
| o3 / o3-pro / o3-mini | Prior | Reasoning; o4-mini noted as succeeded by GPT-5 mini in some copy. |
| o4-mini | Prior | Fast reasoning tier. |
| o1 / o1-pro / o1-mini / o1 Preview | Deprecated | First o-series; previews retired. |
| GPT-4.1 nano / mini | Prior | Smaller GPT-4.1 variants. |
| GPT-4o / GPT-4o mini | API | Omni multimodal workhorses. |
| GPT-4o Search Preview / GPT-4o mini Search Preview | Preview | Web search oriented completions. |
| computer-use-preview | Preview | Computer-use tool specialization. |
| omni-moderation | Current | Text + image safety classification. |
| GPT-4.5 Preview | Deprecated | Large preview model (retired in menus). |
| babbage-002 / davinci-002 | Deprecated | GPT-3 base-class replacements. |
| text-embedding-3-large / small, text-embedding-ada-002 | Embeddings | Embedding API SKUs. |
| text-moderation / text-moderation-stable | Deprecated | Legacy moderation classifiers. |
ChatGPT-branded API SKUs
Marked “for ChatGPT, not recommended for API use” in docs — listed for parity with menus.
| Model | Status | Notes |
|---|---|---|
| GPT-5 Chat / GPT-5.1 Chat / GPT-5.2 Chat / GPT-5.3 Chat | ChatGPT | Instant / chat-tuned variants inside ChatGPT. |
| ChatGPT-4o | Deprecated | GPT-4o flavor used in legacy ChatGPT flows. |
Chat & Completion model IDs (with aliases)
Canonical model id strings and common aliases used by tooling (IDEs, routers). Aliases are shorthand — resolve to the full id in your client. Availability changes; confirm in API model docs.
| Kind | Model ID | Aliases |
|---|---|---|
| OpenAI Chat | gpt-4o |
4o |
| OpenAI Chat | chatgpt-4o-latest |
chatgpt-4o |
| OpenAI Chat | gpt-4o-mini |
4o-mini |
| OpenAI Chat | gpt-4o-audio-preview |
— |
| OpenAI Chat | gpt-4o-audio-preview-2024-12-17 |
— |
| OpenAI Chat | gpt-4o-audio-preview-2024-10-01 |
— |
| OpenAI Chat | gpt-4o-mini-audio-preview |
— |
| OpenAI Chat | gpt-4o-mini-audio-preview-2024-12-17 |
— |
| OpenAI Chat | gpt-4.1 |
4.1 |
| OpenAI Chat | gpt-4.1-mini |
4.1-mini |
| OpenAI Chat | gpt-4.1-nano |
4.1-nano |
| OpenAI Chat | gpt-3.5-turbo |
3.5, chatgpt |
| OpenAI Chat | gpt-3.5-turbo-16k |
chatgpt-16k, 3.5-16k |
| OpenAI Chat | gpt-4 |
4, gpt4 |
| OpenAI Chat | gpt-4-32k |
4-32k |
| OpenAI Chat | gpt-4-1106-preview |
— |
| OpenAI Chat | gpt-4-0125-preview |
— |
| OpenAI Chat | gpt-4-turbo-2024-04-09 |
— |
| OpenAI Chat | gpt-4-turbo |
gpt-4-turbo-preview, 4-turbo, 4t |
| OpenAI Chat | gpt-4.5-preview-2025-02-27 |
— |
| OpenAI Chat | gpt-4.5-preview |
gpt-4.5 |
| OpenAI Chat | o1 |
— |
| OpenAI Chat | o1-2024-12-17 |
— |
| OpenAI Chat | o1-preview |
— |
| OpenAI Chat | o1-mini |
— |
| OpenAI Chat | o3-mini |
— |
| OpenAI Chat | o3 |
— |
| OpenAI Chat | o4-mini |
— |
| OpenAI Chat | gpt-5 |
— |
| OpenAI Chat | gpt-5-mini |
— |
| OpenAI Chat | gpt-5-nano |
— |
| OpenAI Chat | gpt-5-2025-08-07 |
— |
| OpenAI Chat | gpt-5-mini-2025-08-07 |
— |
| OpenAI Chat | gpt-5-nano-2025-08-07 |
— |
| OpenAI Chat | gpt-5.1 |
— |
| OpenAI Chat | gpt-5.1-chat-latest |
— |
| OpenAI Chat | gpt-5.2 |
— |
| OpenAI Chat | gpt-5.2-chat-latest |
— |
| OpenAI Chat | gpt-5.4 |
— |
| OpenAI Chat | gpt-5.4-2026-03-05 |
— |
| OpenAI Chat | gpt-5.4-mini |
— |
| OpenAI Chat | gpt-5.4-mini-2026-03-17 |
— |
| OpenAI Chat | gpt-5.4-nano |
— |
| OpenAI Chat | gpt-5.4-nano-2026-03-17 |
— |
| OpenAI Completion | gpt-3.5-turbo-instruct |
3.5-instruct, chatgpt-instruct |
Wall rows mirror example OpenAI model picker screenshots (Q1 2026 museum snapshot); not a live API guarantee.
References
OpenAI · GPT-4 · ChatGPT · GPT-4o (if available) · API model docs
Compiled for the museum from Wikipedia and OpenAI public posts; not an official OpenAI model sheet.