GPT-6 models: Astra, Sol, Terra and Luna

The GPT-6 lineup

GPT-6 Astra (gpt-6-astra) sits at the top of the GPT-6 models as OpenAI's "most capable model for the most demanding work." For a closer look at the model, see the dedicated GPT-6 Astra guide.

OpenAI followed Astra with four more releases across September and October 2026. GPT-6.1 Sol (gpt-6.1-sol) and GPT-6.1 Terra (gpt-6.1-terra) are the newer additions from September 29 and October 26, 2026, while GPT-6 Luna (gpt-6-luna) is the low-cost model for high-volume jobs and GPT-6 Sol (gpt-6-sol) is the earlier Sol release from September 22, 2026.

Here is how the five models line up by API ID, release date, intended use, and Standard pricing for prompts up to 272K input tokens.

Shared specs and key differences

All five models in the family share the same core limits and modalities. You get a 1,050,000-token context window with a 922,000 max input limit and 128,000 max output tokens, plus text and image input, text output, reasoning tokens, and no audio or video support.

• On the OpenAI API, all five models support the Chat Completions, Responses, and Batch endpoints, while Realtime, Assistants, fine-tuning, and embeddings are not supported.
• Tool calling on gpt-6-astra, gpt-6.1-sol and gpt-6.1-terra requires the Responses API, and gpt-6-luna and gpt-6-sol only allow function calling in Chat Completions when reasoning_effort is set to none.

The table below shows where the five models differ on reasoning effort settings, knowledge cutoffs, and cached input pricing.

API pricing rules

• Prompts over 272K input tokens are billed at 2x the standard input and cache rates and 1.5x the output rate across the entire request. That brings gpt-6-astra to $20 per 1M input tokens and $75 output, gpt-6.1-sol to $4 and $15, gpt-6.1-terra to $2 and $7.50, and gpt-6-luna to $0.20 and $0.75.
• Batch and Flex processing both run at 50% below Standard API rates.
• Fast mode, which OpenAI renamed from Priority processing on July 30, 2026, costs 2x Standard rates.
• Ultrafast mode costs 6x Standard rates and is only available on gpt-6-astra ($60 input, $300 output per 1M tokens) and gpt-6.1-sol ($12 input, $60 output).
• Cache writes are billed at 1.25x the model's uncached input rate with 30-minute retention and 2.5x with 12-hour retention.
• Regional processing endpoints for data residency add a 10% surcharge, which applies to all models released on or after March 5, 2026.

Which GPT-6 model to use

• Use gpt-6-astra for your hardest, most demanding tasks when you can justify $10 per 1M input tokens and $50 per 1M output tokens.
• Pick gpt-6.1-sol for complex coding, computer use, and professional work where you want near-Astra performance at $2 in and $10 out per 1M tokens, which is one fifth of Astra's price.
• Go with gpt-6.1-terra for visual reasoning, precise computer use, and clear writing. It costs $1 in and $5 out per 1M tokens (half the price of gpt-6.1-sol), and in OpenAI's computer-use testing it completed tasks over 3x faster than both gpt-6-astra and gpt-6.1-sol.
• Use gpt-6-luna for focused, high-volume tasks where keeping costs down to $0.10 per 1M input tokens and $0.50 output matters most.
• Keep gpt-6-sol mainly for existing apps that are already pinned to it, since gpt-6.1-sol has the exact same $2 input and $10 output price but cuts cached input in half ($0.10 per 1M tokens vs $0.20 on gpt-6-sol).

GPT-6 in ChatGPT and Codex

Here is how OpenAI maps the GPT-6 models across ChatGPT Chat, ChatGPT Work, and Codex by subscription tier.

• ChatGPT Plus users get Medium and High reasoning in Chat, without access to Extra High or Pro reasoning modes.
• Using gpt-6-astra in the terminal requires Codex CLI 0.153.0 or later, and gpt-6.1-terra requires 0.159.0 or later.

Codex usage limits by model

The table below lists the local message allowances per five-hour window for ChatGPT Plus ($20 a month) and Standard Business plans.

• Pro plans currently have no five-hour message cap, though local messages and cloud tasks share your plan's allowance and weekly limits may still apply; see our Codex usage limits guide for the full breakdown.
• In Codex and ChatGPT Work, gpt-6.1-sol defaults to a 500K context window, and usage drawn from a subscription never triggers the 272K long-context surcharge.

Use GPT-6 through your ChatGPT plan

With Plan2API's ChatGPT integration, you connect your ChatGPT account using a one-time device code at plan2api.com/connect. Plan2API reads the Codex model list live from OpenAI and exposes each model with an openai/<model> prefix (or codex/<model>), such as openai/gpt-6.1-sol, openai/gpt-6-astra, and openai/gpt-6-luna, on a standard OpenAI Chat Completions endpoint (POST /v1/chat/completions) at https://api.plan2api.com/v1. Requests run on your own ChatGPT plan, so pick a model your plan includes.

Plan2API supports streaming, function calling, and reasoning_effort (or OpenRouter-style reasoning.effort) right in the Chat Completions format on Codex models, so you don't have to rewrite your code for the Responses API just to call tools. Every request counts against your ChatGPT subscription's own usage limits, and Plan2API never bills per token.

curl https://api.plan2api.com/v1/chat/completions \
  -H "Authorization: Bearer $PLAN2API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "openai/gpt-6.1-sol", "reasoning_effort": "high", "messages": [{"role": "user", "content": "Refactor this function and explain the change"}]}'