GPT-5.6 API Pricing Guide 2026: Sol, Terra & Luna Token Costs

OpenAI released the GPT-5.6 family on July 9, 2026. It has three API tiers: GPT-5.6 Sol for frontier capability, GPT-5.6 Terra for a balance of intelligence and cost, and GPT-5.6 Luna for fast, cost-sensitive workloads.

The short answer is:

Model	API model ID	Input	Cached input read	Output
GPT-5.6 Sol	gpt-5.6-sol	$5.00	$0.50	$30.00
GPT-5.6 Terra	gpt-5.6-terra	$2.50	$0.25	$15.00
GPT-5.6 Luna	gpt-5.6-luna	$1.00	$0.10	$6.00

Prices are USD per 1 million tokens for standard short-context processing. The gpt-5.6 alias routes to gpt-5.6-sol.

GPT-5.6 Models At A Glance

All three models have a 1,050,000-token context window, 128,000 maximum output tokens, and a February 16, 2026 knowledge cutoff. They accept text and images as input and return text. Streaming, function calling, structured outputs, and reasoning tokens are supported; fine-tuning is not currently supported.

Model	Positioning	Context	Max output	Reasoning effort
GPT-5.6 Sol	Hard professional, coding, research, and agent work	1.05M	128K	none through max
GPT-5.6 Terra	Everyday production work with lower cost	1.05M	128K	none through max
GPT-5.6 Luna	High-volume routing and lighter workloads	1.05M	128K	none through max

OpenAI also added Programmatic Tool Calling, persisted reasoning, explicit prompt-cache controls, and Multi-agent orchestration in beta for the Responses API. Pro mode does not use a separate gpt-5.6-pro model ID. Keep the selected GPT-5.6 model and set reasoning.mode to pro.

Standard And Long-Context Pricing

GPT-5.6 introduces a separately billed cache-write operation. A cache read receives a 90% input discount, while writing tokens into the cache costs 1.25 times the uncached input rate.

For prompts with more than 272K input tokens, OpenAI applies long-context rates to the entire request: input, cached reads, and cache writes cost 2 times their short-context rates, while output costs 1.5 times the short-context rate.

Model	Context band	Input	Cached read	Cache write	Output
GPT-5.6 Sol	Up to 272K input	$5.00	$0.50	$6.25	$30.00
GPT-5.6 Sol	Above 272K input	$10.00	$1.00	$12.50	$45.00
GPT-5.6 Terra	Up to 272K input	$2.50	$0.25	$3.125	$15.00
GPT-5.6 Terra	Above 272K input	$5.00	$0.50	$6.25	$22.50
GPT-5.6 Luna	Up to 272K input	$1.00	$0.10	$1.25	$6.00
GPT-5.6 Luna	Above 272K input	$2.00	$0.20	$2.50	$9.00

All prices are per 1 million tokens. Tool calls can add separate charges. For example, OpenAI currently bills web search per call plus the model tokens consumed by search content.

Batch, Flex, And Priority Prices

OpenAI publishes four processing modes for GPT-5.6. Batch and Flex both reduce short-context token prices by 50%. Priority doubles the standard GPT-5.6 rates.

Model	Processing mode	Input	Cached read	Cache write	Output
GPT-5.6 Sol	Standard	$5.00	$0.50	$6.25	$30.00
GPT-5.6 Sol	Batch / Flex	$2.50	$0.25	$3.125	$15.00
GPT-5.6 Sol	Priority	$10.00	$1.00	$12.50	$60.00
GPT-5.6 Terra	Standard	$2.50	$0.25	$3.125	$15.00
GPT-5.6 Terra	Batch / Flex	$1.25	$0.125	$1.5625	$7.50
GPT-5.6 Terra	Priority	$5.00	$0.50	$6.25	$30.00
GPT-5.6 Luna	Standard	$1.00	$0.10	$1.25	$6.00
GPT-5.6 Luna	Batch / Flex	$0.50	$0.05	$0.625	$3.00
GPT-5.6 Luna	Priority	$2.00	$0.20	$2.50	$12.00

These are short-context rates. The pricing page also lists higher long-context Batch and Flex prices. Regional processing for eligible models released on or after March 5, 2026 has a 10% uplift.

Simple Monthly Cost Example

Suppose an application uses 2 million uncached input tokens and 500,000 output tokens in a month, with standard short-context processing and no paid tool calls.

Model	Input cost	Output cost	Total
GPT-5.6 Sol	2 x $5.00 = $10.00	0.5 x $30.00 = $15.00	$25.00
GPT-5.6 Terra	2 x $2.50 = $5.00	0.5 x $15.00 = $7.50	$12.50
GPT-5.6 Luna	2 x $1.00 = $2.00	0.5 x $6.00 = $3.00	$5.00

For repeated prefixes, cache economics depend on both the first write and later reads. Writing a 1M-token prefix once and reading it nine times costs $10.75 on Sol, compared with $50.00 for ten uncached input passes. The same pattern costs $5.375 instead of $25.00 on Terra, or $2.15 instead of $10.00 on Luna. Output and tool charges are additional.

Use the AI Model Pricing Calculator to test your own input, output, and cache mix.

Which GPT-5.6 Model Should You Choose?

Choose GPT-5.6 Sol when a successful result is worth substantially more than the token bill: difficult coding, research, cybersecurity, science, design, or multi-step professional workflows.

Choose GPT-5.6 Terra as the first production baseline when you need strong reasoning but cannot justify Sol on every request. OpenAI describes Terra as the GPT-5.6 equivalent of the earlier mini tier, although its $2.50/$15 standard price is higher than older mini models.

Choose GPT-5.6 Luna for classification, extraction, routing, lighter agents, and high-volume tasks. It is the cheapest GPT-5.6 option, but older GPT-5 and GPT-5.4 nano-class models can still cost less per token.

For migrations from GPT-5.5 or GPT-5.4, OpenAI recommends testing the same reasoning effort and one level lower. GPT-5.6 may preserve quality with fewer output tokens, but you should validate that claim on your own eval set.

API Example

OpenAI recommends the Responses API for reasoning, tool use, and multi-turn workflows:

from openai import OpenAI

client = OpenAI()

response = client.responses.create(
    model="gpt-5.6-terra",
    reasoning={"effort": "medium"},
    input="Compare these two API designs and return the three highest-risk tradeoffs."
)

print(response.output_text)

Use gpt-5.6-sol when capability is the priority, gpt-5.6-terra for balanced production traffic, or gpt-5.6-luna for efficient high-volume work.

Does GPT-5.6 Remove Codex Weekly Limits?

No official source supports a blanket claim that GPT-5.6 removed weekly limits for individual Codex subscriptions. OpenAI’s current Codex pricing page lists shared five-hour usage windows for local messages and cloud tasks and says that additional weekly limits may apply. The GPT-5.6 ChatGPT help page also says the model uses existing plan allowances.

OpenAI is separately replacing old Enterprise and Edu administrator-configured weekly limits with monthly usage controls. That workspace billing migration does not mean Plus, Pro, or Business usage is unlimited, and it is separate from API RPM and TPM rate limits.

Bottom Line

GPT-5.6 Sol, Terra, and Luna share the same 1.05M context window and 128K maximum output, but their token prices follow a simple 5:2.5:1 input ratio and a 30:15:6 output ratio. Terra is the practical balanced route, Luna is the economical high-volume route, and Sol is the quality-first route.

Budget cache writes separately, watch the 272K long-context threshold, and use Batch or Flex when the workload can tolerate their processing characteristics. See the broader OpenAI API pricing guide and AI API pricing comparison for older OpenAI models and cross-provider alternatives.

Official OpenAI sources checked July 14, 2026:

• OpenAI API pricing
• GPT-5.6 Sol model page
• GPT-5.6 Terra model page
• GPT-5.6 Luna model page
• GPT-5.6 model guidance
• GPT-5.6 launch announcement
• GPT-5.6 in ChatGPT
• Codex pricing and usage limits
• Enterprise and Edu usage-limit migration