OpenAI model catalog adds GPT-5.6 pricing tiers

Teams choosing an OpenAI model now have a clearer split between top-end reasoning, balanced cost, and high-volume workloads. The updated catalog on OpenAI Developers puts GPT-5.6 Sol, Terra, and Luna front and center.

OpenAI updated its model catalog with GPT-5.6 tiers and shared multimodal support.

What changed

The docs now organize the main recommendation around three GPT-5.6 variants: Sol for complex reasoning and coding, Terra for a middle ground on cost and intelligence, and Luna for cost-sensitive, high-volume work. OpenAI also says all latest models accept text and image input, return text output, and support multilingual use and vision.

The catalog ties those models to the Responses API and OpenAI client SDKs, which keeps model selection and integration in the same workflow. That matters for teams comparing models before wiring them into production apps.

• GPT-5.6 Sol: $5 input, $30 output per MTok
• GPT-5.6 Terra: $2.50 input, $15 output per MTok
• GPT-5.6 Luna: $1 input, $6 output per MTok
• Shared limits: 1.05M-token context, 128K max output

Why it matters

For developers, the update makes tradeoffs easier to read at a glance. A single catalog now shows which model fits code-heavy work, which one reduces spend, and which one is tuned for throughput without making teams hunt through separate docs.

It also signals how OpenAI wants teams to think about model choice: not just by capability, but by workload and price band. That is useful for products that need vision input, multilingual output, or tool use without changing integration patterns.

The practical question is no longer whether the model can handle text and images. It is which GPT-5.6 tier best matches latency, budget, and output needs for the next deployment.