OpenAI API Pricing 2026: GPT-5.6, GPT-5.5, GPT-5.4 and Codex Costs
Current OpenAI API prices per 1M tokens for GPT-5.6 Sol, Terra, Luna, GPT-5.5, GPT-5.4 and Codex, including cache writes, long context, Batch, Flex and Priority.
OpenAI’s current API lineup is led by the GPT-5.6 family: Sol for frontier capability, Terra for balanced production work, and Luna for cost-sensitive volume. The three models launched on July 9, 2026 and share a 1.05M-token context window and 128K maximum output.
Current OpenAI API Prices
Standard prices are USD per 1 million tokens.
| Model | Input | Cached read | Cache write | Output | Context |
|---|---|---|---|---|---|
| GPT-5.6 Sol | $5.00 | $0.50 | $6.25 | $30.00 | 1.05M |
| GPT-5.6 Terra | $2.50 | $0.25 | $3.125 | $15.00 | 1.05M |
| GPT-5.6 Luna | $1.00 | $0.10 | $1.25 | $6.00 | 1.05M |
| GPT-5.5 | $5.00 | $0.50 | - | $30.00 | 1.05M |
| GPT-5.5 Pro | $30.00 | - | - | $180.00 | 1.05M |
| GPT-5.4 | $2.50 | $0.25 | - | $15.00 | 1.05M |
| GPT-5.4 mini | $0.75 | $0.075 | - | $4.50 | 400K |
| GPT-5.4 nano | $0.20 | $0.02 | - | $1.25 | 400K |
| GPT-5.3-Codex | $1.75 | $0.175 | - | $14.00 | 400K |
gpt-5.6 is an alias for gpt-5.6-sol. OpenAI does not publish a separate gpt-5.6-pro model ID; Pro is a reasoning mode on Sol.
GPT-5.6 Long-Context Pricing
Requests above 272K input tokens are billed at 2x input and 1.5x output for the entire request.
| Model | Long input | Long cached read | Long cache write | Long output |
|---|---|---|---|---|
| GPT-5.6 Sol | $10.00 | $1.00 | $12.50 | $45.00 |
| GPT-5.6 Terra | $5.00 | $0.50 | $6.25 | $22.50 |
| GPT-5.6 Luna | $2.00 | $0.20 | $2.50 | $9.00 |
GPT-5.6 also changes prompt-caching economics. Cache reads retain the 90% discount, while explicit cache writes are billed at 1.25x normal uncached input. See the full GPT-5.6 pricing guide for cache examples.
Batch, Flex, and Priority
| Processing mode | Price relative to Standard | Use case |
|---|---|---|
| Batch | 50% | Asynchronous jobs with a 24-hour window |
| Flex | 50% | Lower-priority work that accepts variable latency |
| Standard | 100% | Normal interactive traffic |
| Priority | 200% for GPT-5.6 | Latency-sensitive production traffic |
For example, GPT-5.6 Terra costs $1.25/M input and $7.50/M output in Batch or Flex. Priority costs $5/M input and $30/M output. Regional processing adds a 10% uplift for eligible models released on or after March 5, 2026.
Monthly Cost Examples
These examples use 3M input and 1.5M output tokens at short-context Standard rates, with no cache hits:
| Model | Monthly cost |
|---|---|
| GPT-5.4 nano | $2.48 |
| GPT-5.6 Luna | $12.00 |
| GPT-5.4 | $30.00 |
| GPT-5.6 Terra | $30.00 |
| GPT-5.5 | $60.00 |
| GPT-5.6 Sol | $60.00 |
Terra is priced exactly like GPT-5.4, and Sol is priced exactly like GPT-5.5. The upgrade decision should therefore be driven by task success, token efficiency, and migration risk rather than headline token price alone.
Which OpenAI Model Should You Use?
- GPT-5.6 Sol: difficult coding, research, cybersecurity, complex tool use, and high-impact agent tasks.
- GPT-5.6 Terra: default production coding, assistants, document work, and balanced agent workflows.
- GPT-5.6 Luna: extraction, classification, repository search, summarization, and high-volume routing.
- GPT-5.4 mini/nano: keep for mature lower-cost routes when migration to GPT-5.6 has not been evaluated.
- GPT-5.3-Codex: the current official rate card still uses it for Codex code review.
For a deeper routing decision, read GPT-5.6 Sol vs Terra vs Luna.
Codex Limits Are Separate From API Rate Limits
Direct API usage is pay-as-you-go and subject to organization/project RPM and TPM limits. ChatGPT-plan Codex usage is converted into credits and uses product allowances. OpenAI’s current Codex page still describes a five-hour window and says additional weekly limits may apply.
Enterprise/Edu administrators are separately moving from an old weekly workspace credit-control setting to monthly Usage limits. That admin migration does not mean every personal Codex weekly allowance disappeared. See the GPT-5.6 Codex limits guide.
Cost Control Checklist
- Route routine work to Luna and normal production work to Terra before escalating to Sol.
- Keep reusable prompt prefixes stable to increase cache reads.
- Budget explicit cache writes separately on GPT-5.6.
- Treat requests above 272K input tokens as a separate cost class.
- Use Batch or Flex whenever immediate response time is unnecessary.
- Read current RPM, TPM, and usage limits from your OpenAI dashboard.
Use the AI Model Pricing Calculator for your own token mix.
Official Sources
Related Posts
GPT-5.6 API Pricing Guide 2026: Sol, Terra & Luna Token Costs
2026-07-14
AI Coding Agent Cost Comparison 2026: Codex, Claude Code, Cursor, DeepSeek & GPT-5.6
2026-05-07
GPT-5.6 in Codex: Credit Pricing, Multi-Agent, and Weekly Limit Facts
2026-07-14