gpt-6-luna
Our most efficient model for focused, high-volume tasks.
Modalities
In / Out Price
40% off$0.06/ 1M
$0.1 Official price
$0.3/ 1M
$0.5 Official price
Context
1.05M
Released
Sep 22, 2026
API call examples
POST /v1/chat/completionscurl --request POST 'https://api.tokatlas.com/v1/chat/completions' \
--header "Authorization: Bearer $UNIROUTER_API_KEY" \
--header 'Content-Type: application/json' \
--data '{
"model": "gpt-6-luna",
"messages": [
{
"role": "user",
"content": "Explain quantum entanglement."
}
],
"temperature": 0.7
}'Providers & routes
Current provider · OpenAI
Platform configured price · Group: codex
USD / 1M tokens. Each row shows input and output from the same tier. Displayed rates use the lowest available group multiplier; your API key group may differ.
GPT-6 Luna for repeatable text workflows
A candidate for repeatable conversations, classification, and extraction. Validate the output shape and error handling on a small batch first.
Repeated tasks
Batch classification, summarization, and text transformation with common fields.
Long inputs
Provide reference specs and label definitions alongside the working data.
Tunable reasoning
Use format accuracy to compare settings and handle complex cases separately.
Structured responses
Define how to represent unknown values and missing fields so the application can validate them.
About
Repeatable text work
Luna suits classification, text rewriting, and extraction with predictable acceptance rules. Return fixed fields for application validation and send unusual cases to a model chosen for deeper reasoning.
Comparison with GPT-5.6 Luna and Sol
Compare on the same labeling, summary, and formatting cases. Use Luna for high-frequency steps and benchmark Sol for work involving more dependencies and revisions. Establish a lower-effort baseline for simple calls.
Switching an existing workflow
Freeze fields, missing-value conventions, and truncation handling. Test a small rollout for parser compatibility and measure latency, manual corrections, and token use separately.
Judge routine automation by whether extracted fields meet validation rather than by response fluency alone. Run the same post-processing across models on cases with ambiguous or missing fields to identify where manual review is needed.
gpt-6-luna FAQ
When should I start with Luna instead of Sol?
Test Luna on short, easy-to-verify tasks. For work that requires planning across files, run Sol against the same tests and compare completion cost.
Can Luna call tools while reasoning is enabled?
It depends on the endpoint. Before relying on reasoning and tools together, run a minimal test using the Responses example or the exact route you will call.
What tests should I keep when migrating to Luna?
Keep representative prompts and format checks from the previous model. Rerun them with comparable effort and watch short-request latency and retry rate.
Can Luna produce an image file?
Luna delivers text to the application. If the workflow needs an image asset, route that step to a model that generates images.