OpenAI models and products
Which OpenAI models and products are current
In short
OpenAI’s current frontier lineup is GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna. All three are also rolling out in private preview to select enterprise customers in Google Cloud’s Model Garden (Google Cloud). GPT-6.1 Terra launched October 26 at $1 per million input tokens and $5 per million output. OpenAI itself says Sol and GPT-6 Astra remain the better choice for the most complex coding tasks (Terra launch). The Decisions API, built on GPT-6 Luna, is generally available at $0.10 per million input tokens, $0.01 for cached input and nothing for output, with up to 400 questions per call (Decisions API). gpt-6-astra-law costs $10 per million input tokens, $50 output and $1 cached input, which OpenAI labels promotional pricing, available “at least through December 31, 2026” (pricing).
The latest changes are to plans. Starting in December, Plus, Pro and Business usage in ChatGPT and Codex will count against one shared allowance measured in tokens. Instant chat stays outside it (shared pool). OpenAI doubled usage limits on the Pro 100, 200 and 500 tiers and Business Premium, and called off the planned Pro 200 cut (limits). Pro tiers now come with $100, $200 or $500 in monthly API credits (credits). For developers this means Codex and advanced ChatGPT models will draw down the same pool. The API cache lasts 12 hours (caching). Under Usage Policies effective November 20, 2026, a router or AI gateway needs OpenAI’s review and approval to offer access to OpenAI models (policies). Prism, OpenAI’s web app for scientists, shuts down on January 1, 2027 (Prism).
At a glance
| Model or product | What it is | Status | As of | Source |
|---|---|---|---|---|
| GPT-6.1 Terra | Model for high-volume API work; $1 per million input tokens, $5 per million output | Launched October 26 for Plus, Pro, Enterprise and Edu users (Codex changelog also lists Business) and in the API as gpt-6.1-terra | 26 October 2026 | Read |
| GPT-6 Astra | GPT-6 model; API: $10 per million input tokens, $50 output, $1 cached input | Released September 3, in the API a day later | 31 October 2026 | Read |
| GPT-6.1 Sol | GPT-6.1 model | 500k context in Codex and ChatGPT Work | 12 October 2026 | Read |
| GPT-6 Astra for Law (gpt-6-astra-law) | Legal model configuration; $10 per million input tokens, $50 output, $1 cached input | On the API pricing page; promotional pricing, available "at least through December 31, 2026" | 31 October 2026 | Read |
| Decisions API | Built on GPT-6 Luna; returns a probability, a choice or a score; up to 400 questions per call; $0.10 per million input tokens, $0.01 cached input, nothing for output | Generally available; US and EU inference residency | 22 October 2026 | Read |
| ChatGPT and Codex shared usage allowance | One allowance for Plus, Pro and Business usage in ChatGPT and Codex, measured in tokens; Instant chat outside it | Starts in December | 1 November 2026 | Read |
| ChatGPT Pro 100, 200 and 500; Business Premium | ChatGPT plan tiers | Usage limits doubled; planned Pro 200 cut from 20x to 10x the Plus allowance cancelled; extended limits for Dots stay "for now" | 29 October 2026 | Read |
| API credits with ChatGPT Pro and Business | Pro: $100, $200 or $500 per billing period by tier; Business: $20 per paid Standard seat, $100 per paid Premium seat, pooled, capped at $500 per monthly billing period ($6,000 for annual upfront billing) | Available to claim; first grant prorated | 29 October 2026 | Read |
| GPT-6 Astra Pro | Splits one task among parallel agents | 16 October 2026 | Read | |
| OpenAI models on Google Cloud | GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna in Model Garden, bought through Google Cloud Marketplace | Private preview for select enterprise customers; general availability planned in the coming weeks | 21 October 2026 | Read |
| OpenAI on Oracle Marketplace | OpenAI API, Codex and ChatGPT Work through Oracle procurement and eligible Oracle Universal Credits | Live for U.S. Oracle customers | 28 October 2026 | Read |
| Prompt caching | Cache kept forever for logged-in Codex and Work users; 12-hour cache for API developers | Live | 21 October 2026 | Read |
Show all (21)
| Model or product | What it is | Status | As of | Source |
|---|---|---|---|---|
| OpenAI CLI | Official command-line tool for the OpenAI REST API; brew install openai/tools/openai or Go 1.26 or later; images default to gpt-image-2.5-sunburst | Released | 21 October 2026 | Read |
| Open Responses | Spec standardizing agent building blocks such as messages and tool calls | Donated to the Agentic AI Foundation; supported by AWS Bedrock, Databricks, Ollama, OpenRouter and vLLM | 22 October 2026 | Read |
| Rosalind (GPT-Rosalind) | OpenAI's life sciences models and tools, including GPT-Rosalind for reasoning in biology and chemistry | Orchestra Bio's team and technology join Rosalind | 20 October 2026 | Read |
| Daybreak Blue | Sessions on personal accounts | From November 29, content may be used to improve models even with the model-improvement setting off | 31 October 2026 | Read |
| GPT-Live-1.1 | Voice model behind ChatGPT Voice; says "checking..." less | 15 October 2026 | Read | |
| GPT-4o | Retirement in 2027; OpenAI sends users to Luna | 14 October 2026 | Read | |
| Prism | AI web app for scientists, launched in January | Shuts down January 1, 2027 | 30 October 2026 | Read |
| ChatGPT Ads | Ads system in ChatGPT run by a post-trained LLM orchestrator and an auction tool | $1 billion annualized revenue run rate less than 200 days after launch; more than 100,000 advertisers | 21 October 2026 | Read |
| ChatGPT Locked Chats | Hides chats behind a PIN or Face ID | 20 October 2026 | Read |
Milestones
- OpenAI speeds up GPT-6 models by about 50% in ChatGPT.
- GPT-6.1 Sol gets 8x speed at $60 per million output tokens.
- Codex and ChatGPT Work give GPT-6.1 Sol 500k context.
- OpenAI sends GPT-4o users to Luna ahead of a 2027 retirement.
- GPT-Live-1.1 makes ChatGPT Voice say “checking...” less.
- GPT-6 Astra Pro splits one task among parallel agents.
- ChatGPT Locked Chats hide chats behind a PIN or Face ID.
- OpenAI acquires Orchestra Bio, whose team and technology join Rosalind.
- OpenAI ships an official CLI for its REST API.
- Codex and Work users get a permanent cache, API developers a 12-hour cache.
- GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna enter private preview in Google Cloud’s Model Garden.
- The Decisions API exits beta with up to 400 questions per call.
- OpenAI donates its Open Responses spec to the Agentic AI Foundation.
- GPT-6.1 Terra launches at $1 per million input tokens and $5 per million output.
- The OpenAI API, Codex and ChatGPT Work become available in Oracle Marketplace in the U.S.
- ChatGPT Pro and Business get monthly API credits.
- OpenAI doubles usage limits on Pro 100, 200 and 500 and Business Premium, and cancels the Pro 200 cut.
- OpenAI sets a January 1, 2027 shutdown date for Prism.
- gpt-6-astra-law appears on the API pricing page at base Astra rates.
- OpenAI announces one shared ChatGPT and Codex usage pool starting in December.
FAQ
Which OpenAI models are current?
OpenAI is bringing GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna, which it calls its frontier models, to Google Cloud in private preview (Google Cloud). GPT-6.1 Terra launched October 26 in the API as gpt-6.1-terra. OpenAI says Sol and GPT-6 Astra remain the better choice for the most complex coding tasks (Terra launch).
How much does GPT-6.1 Terra cost in the API?
$1 per million input tokens and $5 per million output, 50% below GPT-5.6 Terra and GPT-6.1 Sol. It is available to Plus, Pro, Enterprise and Edu users (the Codex changelog also lists Business) and in the API as gpt-6.1-terra (Terra launch).
How will ChatGPT and Codex usage limits work from December?
Starting in December, Plus, Pro and Business usage in ChatGPT and Codex counts against one shared allowance measured in tokens. Instant chat stays outside it. The balance and reset times will show in Settings → Usage, and once the allowance runs out you can wait for the reset, upgrade or buy credits (shared pool).
Do ChatGPT Pro and Business include OpenAI API credits?
Yes. Pro gets $100, $200 or $500 per billing period, depending on the tier, and the first grant is prorated. Business gets $20 per paid Standard seat and $100 per paid Premium seat, pooled for the workspace and capped at $500 per monthly billing period. Workspaces billed annually upfront get the full year at once, capped at $6,000 (credits).
Is OpenAI Prism shutting down?
Yes. Prism, OpenAI’s AI web app for scientists, launched in January and shuts down on January 1, 2027, according to OpenAI’s Prism pages (Prism).
Latest
- OpenAI’s dots can now read email codes to pass 2FA
- Ramp sees AI token spend grow 1.5% after July’s 42%
- OpenAI trims round goal to $20B after $11B first close
- GPT-6 Astra hits 66.7% on wet-lab tasks, GPT-5.6 Sol 43.8%
- Hacktron gets Daybreak Red access to hunt OpenAI’s bugs
- OpenAI’s CFO keeps data center spending after midterms
- Alfozan pleads not guilty in OpenAI bugging case
- ExploitGym score jumps to 48.5% in OpenAI’s GPT-6 Cyber
- OpenAI’s Brockman calls first half “pretty rough”
- Elastic finds Claude Code and ChatGPT abused in intrusions
