Low-Latency Chat
Power responsive user experiences for routine questions and short conversational turns.
Choose GPT-5.6 Luna for fast, repeatable, cost-conscious text tasks that run frequently or at high volume.
Power responsive user experiences for routine questions and short conversational turns.
Apply labels, routing decisions, and moderation categories across large request volumes.
Convert recurring text formats into fields that are easier for applications to consume.
Produce quick summaries, variants, and first drafts where speed and cost matter.
Best suited to high-volume queues and user-facing flows where fast completion and lower token cost dominate.
Respond quickly to routine product and support questions.
Label and route high-volume text.
Return structured fields from recurring inputs.
Create quick drafts before human or model review.
Test the model in three straightforward steps.
Set the response controls and upload options supported by the model.
Describe the task, add relevant context, and review the streamed response and token usage.
Use the documented endpoint and your API key to bring the same model into your product.
Go from an interactive test to a production integration with predictable controls and usage reporting.
Use the API protocol configured for this model, including streaming where available.
Track input tokens, output tokens, and consumed credits after each response.
Pass the supported generation parameters instead of relying on generic defaults.
Test and operate supported text models through the same AIReiter account and billing system.
Common questions about the online playground, pricing, and API access.
Choose Luna for fast, economical chat, classification, extraction, and first-pass drafting.
Luna prioritizes speed and cost; Terra balances capability, and Sol handles the hardest work.
It is designed for that role, but production quality should still be evaluated on representative requests.
Current input and output token rates are displayed above the playground.
Yes. Follow the linked API documentation and use model ID gpt-5.6-luna.