GPT-6 Sol costs what GPT-5.6 Terra did, and exactly what Claude Sonnet 5.5 does

OpenAI released GPT-6 Sol and GPT-6 Luna on Tuesday, September 22, as the two models below GPT-6 Astra. The developer forum announcement carries the pitch in one line: "50% lower API prices for Sol and Luna compared with GPT‑5.6 promotional pricing."

The rate card bears that out. It also shows that the name Sol now sits on a lower rung than it did a generation ago, and that GPT-6 Sol landed on the same short-context prices Anthropic chose six days later for Claude Sonnet 5.5.

The numbers

Standard processing, dollars per million tokens, from OpenAI's pricing page:

| Model | Input | Cached input | Cache writes | Output | | --- | ---: | ---: | ---: | ---: | | GPT-6 Sol | 2.00 | 0.20 | 2.50 | 10.00 | | GPT-5.6 Sol | 4.00 | 0.40 | 5.00 | 20.00 | | GPT-6 Luna | 0.10 | 0.01 | 0.125 | 0.50 | | GPT-5.6 Luna | 0.20 | 0.02 | 0.25 | 1.20 |

For Sol the 50% holds on every column. For Luna, input halves and output falls further, from $1.20 to $0.50, a cut of about 58%. The baseline is itself a discount: GPT-5.6 Sol's model page says its promotional pricing is "available at least through November 21, 2026."

Same name, lower rung

The GPT-5.6 family had three tiers, and OpenAI's model pages map each to an older name. GPT-5.6 Sol "roughly corresponds to the unsuffixed model tier used in earlier GPT-5 families," and the gpt-5.6 alias routes to it. GPT-5.6 Terra "roughly corresponds to the mini model tier," and Luna to the nano tier.

The GPT-6 guide lists three models too, but with Astra on top: gpt-6-astra "for our highest level of capability," gpt-6-sol "for strong reasoning on demanding tasks." There is no GPT-6 Terra.

OpenAI does not say that GPT-6 Sol replaces GPT-5.6 Terra. Read the rungs instead of the names, though, and the comparison changes. GPT-6 Sol's $2 input matches GPT-5.6 Terra's $2 exactly, and its $10 output is below Terra's $12. At the top of the ladder, GPT-6 Astra lists at $10 input and $50 output, against $4 and $20 for GPT-5.6 Sol. "Half price" is true of the name; by position in the lineup, the middle got a little cheaper and the top got more expensive.

A familiar rate card

Claude Sonnet 5.5, released September 28, lists at $2 input, $10 output, $0.20 for cache reads and $2.50 for cache writes. That is GPT-6 Sol's short-context line to the cent. The match ends at long prompts: on OpenAI's side, "Prompts with more than 272K input tokens are priced at 2x input and cache rates and 1.5x output for the full request," which takes Sol to $4 input and $15 output.

Fine print for integrators

From the two model pages:

• Both models have a 1,050,000-token context window, 922,000 maximum input tokens and 128,000 maximum output tokens.
• Knowledge cutoffs differ: April 20, 2026 for Sol and May 18, 2026 for Luna, so the cheaper model is the more recent one. Both GPT-5.6 models stop at February 16, 2026.
• Both accept reasoning.effort of none, which the GPT-6 guide says Astra does not.
• "Chat Completions supports function calling only with reasoning_effort set to none." Function calling with reasoning switched on means the Responses API.
• EU data residency is available only with Standard processing, and regional processing adds 10% where available. Batch and Flex are half of Standard; Fast mode is double.

OpenAI's announcement page on openai.com, which carries its evaluation claims, refused our fetch, so this story makes no claim about how the models perform.

Primary sources: OpenAI Developer Community announcement, OpenAI API pricing, GPT-6 Sol, GPT-6 Luna, GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.6 Luna, Using GPT-6, Introducing Claude Sonnet 5.5, read 2026-09-29.