GPT-5.6 Luna vs GPT-6.1 Sol
GPT-6.1 Sol is one of OpenAI's newest models and GPT-5.6 Luna an earlier one, so this is also what moving up costs or saves. GPT-5.6 Luna is the cheaper of the 2: for the same tokens, 3 read for each 1 written, GPT-6.1 Sol costs 8.9 times as much. Below are both price lists, what 8 everyday jobs cost a month on each, and how many tokens the same text takes on each, which can change the answer.
- For the same tokens
- GPT-5.6 Luna is cheaper; GPT-6.1 Sol costs 8.9 times as much
- Cheaper to read
- GPT-5.6 Luna, $0.20 against $2
- Cheaper to write
- GPT-5.6 Luna, $1.20 against $10
- Scores higher
- GPT-6.1 Sol, 166 against 156
- Larger window
- The same, 1.05M tokens each
Price per million tokens, side by side
In US dollars, as each maker prints them. The better figure in each row is green.
| GPT-5.6 LunaOpenAI · earlier model | GPT-6.1 SolOpenAI | |
|---|---|---|
| InputTokens the model reads | $0.20 | $2 |
| Cached inputRead again from the cache | $0.02 | $0.10 |
| Cache writePutting text in the cache | $0.25 | $2.50 |
| OutputTokens it writes, thinking included | $1.20 | $10 |
| Blend 3:13 tokens read for each 1 written | $0.45 | $4 |
| Batch inputAnswers collected later | $0.10 | $1 |
| Batch output | $0.60 | $5 |
| Long prompt inputAbove 272K on GPT-5.6 Luna, 272K on GPT-6.1 Sol | $0.40 | $4 |
| Long prompt outputThe standard price where there is no long-prompt price | $1.80 | $15 |
| Context windowTokens it holds at once | 1.05M | 1.05M |
| Capability scoreEpoch Capabilities Index; GPT-5 is 150 | 156 | 166 |
| Longest answerTokens it writes in 1 reply | 128K | 128K |
Price changes
- GPT-5.6 Luna: The price fell by 80%. What changed
What 8 everyday jobs cost a month on each
Each job at its usual number of runs a day, for 30 days, at the standard price. The cheaper of the 2 is green; open any job in the calculator to change it.
| Job | GPT-5.6 Luna | GPT-6.1 Sol | Difference |
|---|---|---|---|
| A chat question300 in, 500 out, 100 a day | $1.980.066 cents a run | $16.800.56 cents a run | 8.5 times |
| A hard question, with thinking300 in, 1,400 out, 100 a day | $5.220.17 cents a run | $43.801.5 cents a run | 8.4 times |
| A long report, summarised60,000 in, 2,000 out, 20 a day | $8.641.4 cents a run | $84.0014 cents a run | 9.7 times |
| A 40-turn agent task2,800,000 in, 40,000 out, 10 a day | $46.3215 cents a run | $364$1.21 a run | 7.8 times |
| A support bot's reply4,000 in, 300 out, 2,000 a day | $48.000.08 cents a run | $4320.72 cents a run | 9 times |
| Sort a message into a category250 in, 5 out, 50,000 a day | $84.000.0056 cents a run | $8250.055 cents a run | 9.8 times |
| Read an invoice into fields1,500 in, 250 out, 1,000 a day | $18.000.06 cents a run | $1650.55 cents a run | 9.2 times |
| Review a pull request25,000 in, 4,500 out, 50 a day | $15.601 cents a run | $1439.5 cents a run | 9.1 times |
Your own job on both
Start from a job, then set its size, how often it runs and whether the answers can wait. Both bills follow every change.
A short question and a few paragraphs back.
about 230 words
Instructions or documents the model has read before cost far less the second time.
about 380 words
A month of this job
- GPT-5.6 Luna$1.98
0.066 cents a run
- GPT-6.1 Sol$16.80
0.56 cents a run
GPT-5.6 Luna saves $14.82 a month; GPT-6.1 Sol costs 8.5 times as much.
This job on every modelWhy the price per token is only the start
Every maker prints 2 main prices: one for the tokens a model reads (input) and one for the tokens it writes (output). On most current models writing costs 3 to 5 times as much as reading, so which model is cheaper can depend on the job. A job that reads a long document and writes a line pays mostly the input price; a chat that writes paragraphs pays mostly the output price. That is why each comparison prices 8 jobs and not one figure.
The same text is not the same number of tokens on every model. Each maker cuts text its own way, and Anthropic says its models from Claude 4.7 on produce about 30 percent more tokens for the same text than earlier ones. A model with the lower price per token can still send the larger bill, so each comparison counts one of our lessons the way both makers count it.
Discounts change the order too. Work that can wait costs half on a batch at the makers that offer one, and text read again from a cache costs a quarter of the input price or less on every current model here. The price table on each comparison shows who offers what.
Thinking is billed as output, and models differ in how much they think and how long they answer. A price list cannot show that. Before you switch, run the same 10 tasks on both models and compare the bills.
Questions
- Which is cheaper, GPT-5.6 Luna or GPT-6.1 Sol?
- GPT-5.6 Luna is cheaper: $0.20 per million tokens read and $1.20 per million written, against $2 and $10 on GPT-6.1 Sol. On a mix of 3 tokens read for each 1 written, GPT-6.1 Sol costs 8.9 times as much.
- Which is better, GPT-5.6 Luna or GPT-6.1 Sol?
- On the Epoch Capabilities Index, an independent score from Epoch AI that combines over 50 benchmarks, GPT-6.1 Sol scores 166 and GPT-5.6 Luna 156, above the top of the range of doubt around GPT-5.6 Luna's score (GPT-5 is 150). One score is an average; test both on your own work before you switch.
- What does a support bot cost a month on GPT-5.6 Luna and GPT-6.1 Sol?
- At 2,000 replies a day, each reading 4,000 tokens (50 percent of them from cache) and writing 300, it costs $48.00 a month on GPT-5.6 Luna and $432 on GPT-6.1 Sol: GPT-6.1 Sol costs 9 times as much for this job.
- What does an AI agent task cost on GPT-5.6 Luna and GPT-6.1 Sol?
- One 40-turn task that reads 2,800,000 tokens in all, 90 percent of them from cache, and writes 40,000, costs 15 cents on GPT-5.6 Luna and $1.21 on GPT-6.1 Sol. At 10 tasks a day that is $46.32 and $364 a month.
- Which has the larger context window, GPT-5.6 Luna or GPT-6.1 Sol?
- Neither: both hold 1.05M tokens at once, about 1,575 pages of English text.
- Does the same text take the same number of tokens on GPT-5.6 Luna and GPT-6.1 Sol?
- Yes, both are counted the same way: our lesson "How AI writes", 1,700 words long, is 2,053 tokens on each.
- Is it cheaper to move from GPT-5.6 Luna to GPT-6.1 Sol?
- No, it costs more per token: GPT-6.1 Sol charges $2 per million tokens read and $10 written, against $0.20 and $1.20 on GPT-5.6 Luna, 8.9 times as much for the same tokens. It pays off only if better answers save more than that, for example in fewer retries.
- Can I pay less on GPT-5.6 Luna and GPT-6.1 Sol for work that can wait?
- Yes, on both. With a batch, where you collect the answers later instead of at once, GPT-5.6 Luna costs $0.10 per million tokens read and $0.60 written, and GPT-6.1 Sol $1 and $5.
Sources