OPEN DATA · UPDATED DAILY
GPT API Pricing: OpenAI Official vs LanoX (Daily Record)
GPT models on LanoX cost up to 83.3% less than OpenAI's list price in the Premium group and up to 87.1% less in the Saver group. Those are the figures on the LanoX model pricing page badges ("At most 83.3% OFF" and "At most 87.1% OFF"), which we read without logging in on 2026-11-03 at 03:58 (UTC+4). For example, gpt-6.1-sol costs $0.334 input and $1.67 output per 1M tokens on LanoX, against $2.00 and $10.00 at OpenAI.
8 GPT models, each with OpenAI's official token price, the LanoX public price and the gap between them. Over 16 days of daily records, the LanoX price changed 0 times in the Premium group and 2 times in the Saver group. No snapshot was saved on 2026-10-18.
Key numbers
How much does each GPT model cost on LanoX vs OpenAI?
All 8 models are listed below, priced per 1M tokens in USD for input up to 272K tokens. OpenAI's prices come from the "Standard" table on the OpenAI API pricing page, short context, checked 2026-11-03. gpt-6-sol, gpt-5.6-terra and gpt-5.6-luna appear in the "All models" list of that same table. LanoX prices are the 1–272K tier on the public LanoX model pricing page, read 2026-11-03.
| Model | Group | Official input | LanoX input | Official output | LanoX output | Input gap | Output gap |
|---|---|---|---|---|---|---|---|
| gpt-6.1-sol | Premium | $2.00 | $0.334 | $10.00 | $1.67 | −83.3% | −83.3% |
| gpt-6-astra | Premium | $10.00 | $1.67 | $50.00 | $8.35 | −83.3% | −83.3% |
| gpt-6-astra | Saver | $10.00 | $1.29 | $50.00 | $6.45 | −87.1% | −87.1% |
| gpt-5.6-sol | Premium | $4.00 | $0.668 | $20.00 | $3.34 | −83.3% | −83.3% |
| gpt-5.6-sol | Saver | $4.00 | $0.516 | $20.00 | $2.58 | −87.1% | −87.1% |
| gpt-6-sol | Premium | $2.00 | $0.334 | $10.00 | $1.67 | −83.3% | −83.3% |
| gpt-6-luna | Premium | $0.10 | $0.0167 | $0.50 | $0.0835 | −83.3% | −83.3% |
| gpt-5.6-terra | Premium | $2.00 | $0.334 | $12.00 | $2.004 | −83.3% | −83.3% |
| gpt-5.6-luna | Premium | $0.20 | $0.0334 | $1.20 | $0.2004 | −83.3% | −83.3% |
| gpt-6.1-terra | Premium | $1.00 | $0.167 | $5.00 | $0.835 | −83.3% | −83.3% |
Gap = (LanoX price − official price) ÷ official price × 100, rounded half up to one decimal place. A negative figure means LanoX is cheaper. Example: gpt-6.1-sol output, (1.67 − 10) ÷ 10 = −83.3%.
What do prompt caching reads and writes cost?
Cache prices follow the same 83.3% (Premium) and 87.1% (Saver) pattern. "Official cache write" is OpenAI's 30-minute cache write price. LanoX lists some very small cache prices to four decimal places, so a few rows land slightly off the group figure. gpt-5.6-luna cache read, for example, is −83.5%, and gpt-6-luna cache read is −83.0%.
| Model | Group | Official cached input | LanoX cache read | Official cache write | LanoX cache write | Cache read gap | Cache write gap |
|---|---|---|---|---|---|---|---|
| gpt-6.1-sol | Premium | $0.10 | $0.0167 | $2.50 | $0.4175 | −83.3% | −83.3% |
| gpt-6-astra | Premium | $1.00 | $0.167 | $12.50 | $2.0875 | −83.3% | −83.3% |
| gpt-6-astra | Saver | $1.00 | $0.129 | $12.50 | $1.6125 | −87.1% | −87.1% |
| gpt-5.6-sol | Premium | $0.40 | $0.0668 | $5.00 | $0.835 | −83.3% | −83.3% |
| gpt-5.6-sol | Saver | $0.40 | $0.0516 | $5.00 | $0.645 | −87.1% | −87.1% |
| gpt-6-sol | Premium | $0.20 | $0.0334 | $2.50 | $0.4175 | −83.3% | −83.3% |
| gpt-6-luna | Premium | $0.01 | $0.0017 | $0.125 | $0.0209 | −83.0% | −83.3% |
| gpt-5.6-terra | Premium | $0.20 | $0.0334 | $2.50 | $0.4175 | −83.3% | −83.3% |
| gpt-5.6-luna | Premium | $0.02 | $0.0033 | $0.25 | $0.0418 | −83.5% | −83.3% |
| gpt-6.1-terra | Premium | $0.05 | $0.0084 | $1.25 | $0.2088 | −83.2% | −83.3% |
What changes when input goes above 272K tokens?
Prices step up above 272K input tokens. OpenAI's pricing page splits text prices into "Short: ≤272K input tokens" and "Long: >272K input tokens", and LanoX lists a separate tier above 272K. Long context has its own tiered prices, so check the LanoX pricing page for current rates. As of 2026-11-03, OpenAI's page shows a long-context price for only 1 of these 8 models: gpt-5.6-sol, at $8.00 input, $0.80 cached input, $10.00 cache write and $30.00 output.
| Model | Group | LanoX input | LanoX cache read | LanoX cache write | LanoX output | OpenAI official (long context), input / output | Input gap | Output gap |
|---|---|---|---|---|---|---|---|---|
| gpt-6.1-sol | Premium | $0.668 | $0.0334 | $0.835 | $2.505 | not found | — | — |
| gpt-6-astra | Premium | $3.34 | $0.334 | $4.175 | $12.525 | not found | — | — |
| gpt-6-astra | Saver | $2.58 | $0.258 | $3.225 | $9.675 | not found | — | — |
| gpt-5.6-sol | Premium | $1.336 | $0.1336 | $1.67 | $5.01 | $8.00 / $30.00 | −83.3% | −83.3% |
| gpt-5.6-sol | Saver | $1.032 | $0.1032 | $1.29 | $3.87 | $8.00 / $30.00 | −87.1% | −87.1% |
| gpt-6-sol | Premium | $0.668 | $0.0668 | $0.835 | $2.505 | not found | — | — |
| gpt-6-luna | Premium | $0.0334 | $0.0033 | $0.0418 | $0.1253 | not found | — | — |
| gpt-5.6-terra | Premium | $0.668 | $0.0668 | $0.835 | $3.006 | not found | — | — |
| gpt-5.6-luna | Premium | $0.0668 | $0.0067 | $0.0835 | $0.3006 | not found | — | — |
| gpt-6.1-terra | Premium | $0.334 | $0.0167 | $0.4175 | $1.2525 | not found | — | — |
What does a typical monthly workload cost?
Take a workload of 10M input tokens and 2M output tokens in the first tier, with no caching. It costs $6.68 on gpt-6.1-sol (Premium), compared with $40.00 at OpenAI's list price. The formula comes from the LanoX pricing FAQ: input tokens ÷ 1,000,000 × input rate, plus output tokens ÷ 1,000,000 × output rate. The figures below are estimated from the 2026-11-03 pricing pages.
| Model | Group | OpenAI official | LanoX | Gap |
|---|---|---|---|---|
| gpt-6.1-sol | Premium | $40.00 | $6.68 | −83.3% |
| gpt-6-astra | Premium | $200.00 | $33.40 | −83.3% |
| gpt-6-astra | Saver | $200.00 | $25.80 | −87.1% |
| gpt-6-luna | Premium | $2.00 | $0.334 | −83.3% |
Who should not rely on this page?
This page covers 8 models at OpenAI's Standard tier only, so it will not fit every case:
- Flex/Batch, Fast or Ultrafast pricing. OpenAI's page has those 3 other tiers. This page compares the Standard tier only.
- Official long-context comparison for most models. As of 2026-11-03, OpenAI's page lists a >272K price for only 1 of the 8 models here (gpt-5.6-sol).
- A fixed official price for gpt-5.6-sol after 2026-11-21. OpenAI calls that price promotional "at least through November 21, 2026".
- Negotiated or enterprise prices. Every price here is a public list price, read without logging in on 2026-11-03. Contract prices are not covered.
How often has the LanoX price changed?
Over 16 days of records since 2026-10-17, the LanoX GPT price changed 0 times in the Premium group and 2 times in the Saver group. We save the LanoX public price for each model once a day. The CSV has one row per model per day, and a change is counted whenever the input or output price differs from the previous saved day, counted separately for each group. No snapshot was saved on 2026-10-18; those days are left out, not filled in. Models added later, such as gpt-6.1-terra, have fewer days on record. Per-model counts are in the CSV.
How were these numbers collected?
- We read the public LanoX model pricing page once every 24 hours. No login, no coupons, no enterprise prices.
- Official prices come from OpenAI's public API pricing page (Standard tier) and are re-checked by hand. The last check was 2026-11-03.
- The main table uses the first tier (input up to 272K tokens), and the tier above 272K is listed separately. Prices exclude long-context and other surcharges unless the table says otherwise.
- gpt-5.6-sol: OpenAI's page says "GPT-5.6 Sol’s promotional pricing is available at least through November 21, 2026." The official price may change after that date, so check OpenAI's page for the current figure.
- Gaps are calculated from the two prices in the table with exact decimal arithmetic, then rounded half up to 1 decimal place.
- Disclosure: LanoX publishes this page and is 1 of the 2 prices shown. Raw daily snapshots are kept so any figure can be checked.
Sources: OpenAI API pricing · LanoX model pricing
Related: LLM API price gap index (LanoX vs OpenRouter) · Claude API pricing (coming soon) · Codex setup guide · LanoX docs
What else do readers ask? (FAQ)
Why are there two price groups?
LanoX lists GPT models in 2 groups, Premium and Saver, and each group has its own price list. On 2026-11-03, Premium covers all 8 models above, while Saver covers 2 of them (gpt-6-astra and gpt-5.6-sol). The badges read "At most 83.3% OFF" for Premium and "At most 87.1% OFF" for Saver. Where a model appears in both groups, the table shows both prices.
What does the 272K tier mean?
272K is the input size of a single request. OpenAI's pricing page labels short context "≤272K input tokens" and long context ">272K input tokens", and LanoX lists a 1–272K tier plus a tier above 272K. Long context has its own tiered prices; see the pricing page.
Which price do I actually pay?
You pay the LanoX rate. gpt-6.1-sol output in Premium, for instance, is charged at $1.67 per 1M tokens, and the $10.00 OpenAI price is not added on top. From the LanoX pricing page FAQ: "You pay the LanoX rate for that model. The provider's published price is shown for comparison; the two prices are not added together."
Do I need a subscription, or do I pay for a full 1M tokens?
No to both. The LanoX pricing FAQ says LanoX is "pay-as-you-go, with no monthly fee or subscription", and that "Per million tokens is a pricing unit, so you pay for the tokens you use." For example, 1 gpt-6.1-sol request in Premium with 2,000 input tokens and 500 output tokens costs $0.001503 at the 2026-11-03 rates.
How do I find the model ID to call?
The 8 names in the first column are the model IDs shown on the LanoX pricing page. The LanoX pricing FAQ adds that you can "use your API key to call GET /v1/models for the model IDs available to your account". Setup steps are in the LanoX docs.
What has been corrected? (change log)
- 2026-11-03: First published. No data corrections so far.