OpenAI GPT models

Choose the right model in the family

Start with the workload and service target, then pick the smallest model that passes a representative evaluation set.

• Choose Sol for the hardest coding, research and long-running professional workflows.
• Choose Terra when quality still matters but the workload needs a lower unit cost.
• Choose Luna for classification, extraction, routing and other high-volume work after task-specific evaluation.

Inputs, outputs, context and tools

• Text and image input with text output across the current GPT-5.6 family.
• A 1.05M context window and up to 128K output support large repositories and document sets.
• Function calling, web search, file search and computer use are published tool options.
• Reasoning can be adjusted from none to max so latency and depth can be tuned per task.

Where this family fits

These are strong starting points—not universal recommendations. Test them with your own prompts, tools and acceptance criteria.

• Repository-scale coding and code review.
• Long-document synthesis with citations verified by your application.
• Tool-using agents that need structured calls and recovery logic.
• High-volume transformation and extraction on Terra or Luna.

Reasoning, context and cost controls

• Start with medium reasoning, then increase it only when the evaluation set shows a quality gain.
• Trim repeated conversation history even when the context window is large.
• Record latency, input, output, tool calls and retries per completed task—not per isolated request.

Send a minimal request

This example uses the first provider model ID in the table. If Model Plaza shows a different ID for your API key, copy the Model Plaza value exactly.

curl https://app.llmfly.ai/v1/chat/completions \
  -H "Authorization: Bearer $LLMFLY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "messages": [{"role": "user", "content": "Explain this request in three bullets."}],
    "stream": false
  }'

Limits and compatibility notes

• A published capability does not guarantee that every LLMFly AI route exposes the same endpoint or tool.
• Large context can raise latency and cost; it is not a substitute for retrieval and context selection.
• Keep model IDs in configuration and test a fallback before production rollout.

Production evaluation checklist

• Run 20–50 representative tasks and define pass criteria before testing.
• Measure time to first token, total latency, input and output tokens, tool retries, and cost per successful task.
• Test the endpoint, streaming mode, tool schemas, and structured output used by the application.
• Verify 400, 401, 404, 429, and temporary 5xx handling with bounded retries.
• Keep the model ID in configuration and validate a fallback before moving production requests.

What to confirm in Model Plaza

Model Plaza shows which model IDs your API key can use, their billing multipliers, availability, supported endpoints, and compatibility notes. Confirm these details with a small request before deployment.

• All three accept text and image input and return text.
• Reasoning effort ranges from none through max.
• OpenAI lists functions, web search, file search and computer use among supported tools.