OpenAI GPT 5.6 Luna: AI Large Language Models
Generate text fast with OpenAI's GPT 5.6 Luna. Cost-effective multimodal reasoning for classification and screening. No subscription.
Generate text fast with OpenAI's GPT 5.6 Luna. Cost-effective multimodal reasoning for classification and screening. No subscription.
About this model
GPT 5.6 Luna is OpenAI's fast, cost-effective member of the GPT 5.6 multimodal reasoning family, tuned for high-volume, low-latency workloads. It accepts mixed text and image inputs, supports a system prompt to steer behavior, and can call an integrated web search tool for real-time information. Reasoning depth is adjustable via the reasoning_effort parameter, letting you trade latency for thoroughness on a per-request basis. Luna is the right starting point for automated classification, content screening, and simple extraction pipelines that need to run at scale without the cost of the larger Terra or Sol models.
Cost analysis
| Provider | Cost | Notes |
|---|---|---|
| muapiapp | $2.00/M input, $12.00/M output | Token-based billing. Minimum $0.0002 per call. |
| Fal.ai | Not available | No dedicated GPT 5.6 Luna endpoint at this time. |
| Replicate | Not available | No dedicated GPT 5.6 Luna endpoint at this time. |
Token-based billing. Minimum $0.0002 per call.
No dedicated GPT 5.6 Luna endpoint at this time.
No dedicated GPT 5.6 Luna endpoint at this time.
** Competitor pricing is estimated based on similar model architectures and usage tiers.
Configuration schema
| Parameter | Type | Description | Default |
|---|---|---|---|
| Prompt | string | The user message or instruction. | Summarize this article in three bullet points. |
| Image URL | string | Optional image URL for multimodal requests. | undefined |
| System Prompt | string | Optional system-level instruction to guide model behavior. | You are an expert tutor. |
| Web Search Switch | boolean | Enable web search tool for real-time information retrieval. | false |
| Reasoning Effort | Enum (4 options) | Level of reasoning depth: low, medium, high, or xhigh. | low |
The user message or instruction.
Summarize this article in three bullet points.Optional image URL for multimodal requests.
undefinedOptional system-level instruction to guide model behavior.
You are an expert tutor.Enable web search tool for real-time information retrieval.
falseLevel of reasoning depth: low, medium, high, or xhigh.
lowDeveloper documentation
POST /api/v1/gpt-5-6-luna — returns request_id, poll for result via /api/v1/predictions/{id}/result.
POST /api/v1/gpt-5-6-luna/stream — returns a live Server-Sent Events stream (Content-Type: text/event-stream). Each chunk has the format data: {"choices":[{"delta":{"content":"text"}}]}, ending with data: [DONE].
curl -X POST "https://api.muapi.ai/api/v1/gpt-5-6-luna" \
-H "Authorization: Bearer $YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "Classify this support ticket as billing, technical, or general.",
"reasoning_effort": "low"
}'
Frequently asked
Luna is optimized for speed and efficiency, making it ideal for high-volume automated classification, screening, and simple extraction tasks where low latency and cost matter more than deep reasoning.
Pricing is token-based at $2.00 per million input tokens and $12.00 per million output tokens, with a minimum charge of $0.0002 per call.
Yes, Luna supports mixed text and image (image_url) inputs in a single message.
You can control the depth of reasoning by passing a reasoning_effort value of 'low', 'medium', 'high', or 'xhigh'. Higher levels provide more thorough logic but increase latency — Luna is typically run at 'low' for maximum speed.
Use the streaming endpoint (/gpt-5-6-luna/stream) for interactive chat interfaces where the user expects real-time feedback. Use the async endpoint (/gpt-5-6-luna) for background tasks, workflows, and automated scripts.
Upgrade to Terra for balanced business and analytical tasks, or to Sol for complex math, programming, and scientific reasoning where Luna's speed-optimized responses aren't thorough enough.