GPT-6 Sol, Luna and GPT-6.1 Sol API: Pricing, How They Differ from Astra, Which to Use

OpenAI's GPT-6 series sits in the Onysoft catalog in three price tiers: the flagship GPT-6 Astra, GPT-6 Sol below it, and GPT-6 Luna in the speed and cost tier. GPT-6.1 Sol, an upgrade to Sol, has been added alongside them. All four offer a 1,050,000-token context window and a 128,000-token maximum output; the real difference is price: 1M input tokens cost $15 on Astra, $3 on Sol and $0.15 on Luna.

This guide covers Sol, Luna and GPT-6.1 Sol: a USD and TL price table, the ratios against Astra, the Pro and :batch IDs, the price tier that applies to prompts above 272,000 tokens, and which model fits which job. Values are taken from the Onysoft catalog as of October 5, 2026. For Astra itself, see the GPT-6 Astra API guide.

How the GPT-6 Family Lines Up in the Catalog

According to the maker's catalog descriptions, the series is positioned as follows:

• GPT-6 Astra (openai/gpt-6-astra): the flagship of the series.
• GPT-6 Sol (openai/gpt-6-sol): the cost-efficient high-end model, positioned below Astra and above Luna; described for demanding professional work.
• GPT-6.1 Sol (openai/gpt-6.1-sol): an upgrade to GPT-6 Sol; the description highlights agentic coding, computer use and document-heavy professional work.
• GPT-6 Luna (openai/gpt-6-luna): the fast, cost-efficient model of the series; described for high-volume and latency-sensitive workloads such as chat, classification and lightweight agentic tasks.

The technical frame of all four is the same in the catalog: a 1,050,000-token context window and a 128,000-token maximum output. So the choice is driven by how hard the workload is and by budget, not by the context limit.

Note: as of October 5, 2026 there is no "GPT-6 Terra" model in the catalog; in the Terra tier, GPT-5.6 Terra (openai/gpt-5.6-terra) is active. For the full series, see the GPT models page.

GPT-6 Sol, Luna and GPT-6.1 Sol Pricing: USD and TL Table

The table below shows Onysoft list prices per 1 million tokens:

Model	API ID	Context (tokens)	1M input	1M output	1M input (TL)	1M output (TL)
GPT-6 Astra	openai/gpt-6-astra	1,050,000	$15.00	$75.00	735.52 TL	3,677.61 TL
GPT-6 Sol	openai/gpt-6-sol	1,050,000	$3.00	$15.00	147.10 TL	735.52 TL
GPT-6.1 Sol	openai/gpt-6.1-sol	1,050,000	$3.00	$15.00	147.10 TL	735.52 TL
GPT-6 Luna	openai/gpt-6-luna	1,050,000	$0.15	$0.75	7.36 TL	36.78 TL
GPT-6 Astra (batch)	openai/gpt-6-astra:batch	1,050,000	$7.50	$37.50	367.76 TL	1,838.81 TL
GPT-6 Sol (batch)	openai/gpt-6-sol:batch	1,050,000	$1.50	$7.50	73.55 TL	367.76 TL
GPT-6 Luna (batch)	openai/gpt-6-luna:batch	1,050,000	$0.075	$0.375	3.68 TL	18.39 TL
GPT-5.6 Sol	openai/gpt-5.6-sol	1,050,000	$3.00	$15.00	147.10 TL	735.52 TL
GPT-5.6 Terra	openai/gpt-5.6-terra	1,050,000	$3.00	$18.00	147.10 TL	882.63 TL
GPT-5.6 Luna	openai/gpt-5.6-luna	1,050,000	$0.30	$1.80	14.71 TL	88.26 TL

Prices are Onysoft list prices as of October 5, 2026; see /models for live pricing. TL equivalents use 1 USD = 49.0348 TL (Central Bank of the Republic of Türkiye rate for October 2, 2026); billing uses the current central bank rate at the time of the request.

The ratios in the table:

• Astra is 5 times Sol; Sol is 20 times Luna. For the same token volume, Astra costs 100 times as much as Luna.
• GPT-6.1 Sol and GPT-6 Sol are the same price ($3 per 1M input, $15 per 1M output); GPT-5.6 Sol is listed at the same price too.
• GPT-6 Luna is cheaper than GPT-5.6 Luna: input $0.15 instead of $0.30 (50% lower), output $0.75 instead of $1.80 (58% lower).

Example monthly bill: an application that processes 50 million input tokens and 10 million output tokens a month pays $15 (735.52 TL) on Luna, $300 (14,710.44 TL) on Sol and $1,500 (73,552.20 TL) on Astra. The same volume on GPT-5.6 Luna comes to $33 (1,618.15 TL). Use the cost calculator for your own volume.

The Pro IDs: Same Token Price, Different Token Consumption

The catalog also has IDs with the -pro suffix for Sol, Luna and GPT-6.1 Sol: openai/gpt-6-sol-pro, openai/gpt-6-luna-pro, openai/gpt-6.1-sol-pro. Their list price per 1M tokens is the same as their non-Pro counterparts.

The difference is in how they run: per the catalog description, a Pro ID is the same underlying model served with reasoning.mode set to pro for higher-quality responses on complex tasks. The same description notes that Pro mode spends far more tokens. The upshot: the token price does not change, but the amount per request can go up.

When you try a Pro ID, compare the usage and cost fields of the response against the non-Pro ID; decide on the actual amount and answer quality for the same prompt on both IDs, not on the list price.

Long Prompts: The 272,000-Token Threshold

A long-context price tier applies to GPT-6 Sol, GPT-6 Luna and GPT-6.1 Sol (including the Pro and batch IDs): once a request's input reaches 272,000 tokens, input is billed at 2 times and output at 1.5 times the list price for that request.

Example: a GPT-6 Sol request with 300,000 input tokens and 2,000 output tokens would cost $0.93 without the tier; with the tier it costs $1.845. If you plan to use the 1,050,000-token context to the full, budget with this tier in mind.

In practice, table prices apply to prompts that stay under 272,000 tokens. Splitting very long documents into parts, or retrieving only the relevant sections (RAG), keeps you under the tier and lowers input cost as well; see the AI API cost optimization guide for details.

The :batch IDs and Aliases

Batch: the :batch IDs of Sol and Luna are billed at half the standard list price:

• openai/gpt-6-sol:batch and openai/gpt-6-sol-pro:batch — $1.50 per 1M input, $7.50 per 1M output
• openai/gpt-6-luna:batch and openai/gpt-6-luna-pro:batch — $0.075 per 1M input, $0.375 per 1M output

Batch IDs are for bulk work where the answer does not need to arrive immediately (bulk labeling, evaluation sets, regression tests). Measure response times with your own workload before production use. As of October 5, 2026 there is no active batch ID for GPT-6.1 Sol; you can call 6.1 Sol only through its standard and Pro IDs.

Aliases: ~openai/gpt-sol-latest and ~openai/gpt-luna-latest redirect to the latest model in the respective family (the leading ~ is required). As of October 5, 2026 they are listed in the catalog at the same price as Sol and Luna respectively. If you do not want behavior to change on its own between releases, prefer the fixed ID in production.

Which One When? A Decision Table

The mapping below is based on how the maker positions the models and on the unit prices in the table:

Workload	Where to start	Why
Chatbots, classification, labeling, short summaries	openai/gpt-6-luna	The model the maker describes for high-volume, latency-sensitive work; one twentieth of the Sol price.
Demanding professional work, coding assistance, long analysis	openai/gpt-6-sol	The high-end model of the series at one fifth of the Astra price.
Agentic coding, computer use, document-heavy work	openai/gpt-6.1-sol	The areas the maker highlights for 6.1 Sol; same price as GPT-6 Sol.
A complex one-off task where quality comes first	openai/gpt-6-sol-pro	Same token price, higher token consumption; track the amount through usage.
Non-urgent bulk work	openai/gpt-6-luna:batch or openai/gpt-6-sol:batch	Half the standard price.
The hardest tasks, where budget comes second	openai/gpt-6-astra	The flagship of the series; 5 times the Sol price.

Rule of thumb: start with Luna, escalate the cases where quality falls short to Sol (or 6.1 Sol), and reserve Astra for the steps that really need it. If you want to compare an alternative at the same list price as GPT-6 Sol ($3 / $15), Claude Sonnet 5.5 can be called through the same API with the same key; you can try both with the same prompt in the Playground. To have the model chosen automatically per request, OnyRouter (onysoft/auto) is also available.

First Request with Python and curl

To get started, create an account, generate an API key with the sk-ony- prefix in the dashboard, and add balance; no separate contract with the model maker is needed. The official openai Python package works as is in streaming mode once you change base_url (in non-streaming responses the OpenAI object is in the data field):

from openai import OpenAI

client = OpenAI(
    base_url="https://api.onysoft.com/v1",
    api_key="sk-ony-YOUR_KEY",
)

stream = client.chat.completions.create(
    model="openai/gpt-6-sol",
    messages=[
        {"role": "user", "content": "List the risky clauses in this draft contract by clause number."}
    ],
    stream=True,
)
for chunk in stream:
    if not chunk.choices:
        continue
    delta = chunk.choices[0].delta.content
    if delta:
        print(delta, end="")

For high-volume short jobs, a non-streaming curl request to Luna; capping output with max_tokens keeps cost predictable:

curl https://api.onysoft.com/v1/chat/completions \
  -H "Authorization: Bearer sk-ony-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-6-luna",
    "max_tokens": 200,
    "messages": [
      {"role": "system", "content": "Classify the message with one word: billing, technical, refund, other."},
      {"role": "user", "content": "I was charged twice last month and want one payment back."}
    ]
  }'

Non-streaming responses come back in a success/data envelope; the OpenAI-format response object is in the data field. For streaming, add "stream": true to the body. If you are moving from GPT-5.x, the only thing that has to change is the model name; see the GPT-5.6 guide for details and the API documentation for supported parameters and error codes. The same key reaches all 754 models in the catalog.