Loading page
OpenAI
GPT-6 Luna
GPT-6 Luna is a chat model from OpenAI, tracked here with third-party throughput measurements of roughly 151 output tokens per second and about 1.46 seconds to first token.
Input from
$0.050 / 1M tokens
across 2 providers
API Pricing
Cheapest on OpenRouter — 40% below avg| Provider | Input / 1M | Output / 1M | Cached / 1M |
|---|---|---|---|
| $0.050 | $0.250 | $0.0050 | |
| $0.100 | $0.500 | $0.010 | |
| $0.100 | $0.500 | $0.010 |
Prices updated daily. Last check: Sep 23, 2026
GPT-6 Luna pricing by provider
Input, output, and batch rates, plus alternatives, for one provider at a time.
Performance & Benchmarks
Source: Artificial Analysis →Intelligence
37.3 / 100
Output Speed
167 t/s
Latency (TTFT)
44.7s
Reasoning & Knowledge
- Humanity's Last Exam38.5%
Coding
- SciCode54.6%
Instruction & Long Context
- Long-Context Reasoning83.3%
Benchmarks measured Sep 2026. Scores are independent evaluations, not vendor-reported.
Model Details
General
- Creator
- OpenAI
- Modalities
- Text
Capabilities
- Tool Calling
- No
- Open Source
- No
Strengths & Limitations
Strengths
- Measured output throughput of approximately 150.7 tokens per second in third-party testing, which supports streaming chat interfaces
- Time to first token of roughly 1.46 seconds, giving a predictable initial latency budget for interactive applications
- Distributed as a chat-completion style model, so it slots into existing OpenAI-compatible client code without rearchitecting request handling
- Carries OpenAI's GPT naming, meaning tooling, SDKs, and prompt patterns built for other GPT chat models generally transfer
- Performance figures sourced from Artificial Analysis rather than vendor-published marketing numbers
- Listed across providers on this page, so cost per token can be compared directly against alternative chat models
Limitations
- We do not currently track a confirmed context window size for GPT-6 Luna, so long-document workloads need verification against OpenAI's documentation
- Supported input and output modalities are not recorded in our database — assume text chat unless OpenAI's docs state otherwise
- No standardized reasoning, coding, or knowledge benchmark scores are held in our catalog for this model, making direct quality comparison harder
- A time to first token near 1.5 seconds is noticeable in latency-sensitive voice or autocomplete use cases
- Measured throughput and latency vary by provider and load; the figures shown are averages, not service guarantees
Key Features
•Chat-completion API interface
•Measured output speed of ~150.7 tokens per second (Artificial Analysis)
•Measured time to first token of ~1,460 ms (Artificial Analysis)
•Streaming response support typical of OpenAI chat endpoints
•Available through multiple inference providers tracked on this page
•Part of OpenAI's GPT chat model naming line
•Per-provider price comparison in the live pricing table
About GPT-6 Luna
GPT-6 Luna is a conversational (chat) model released under OpenAI's GPT model naming. Our catalog entry records it as a text chat endpoint offered through inference API providers, and we track it primarily so its serving performance and per-provider pricing can be compared against other chat models listed on this site.
The measurements we hold for GPT-6 Luna come from Artificial Analysis: an average output speed of about 150.7 tokens per second and a time to first token of roughly 1,460 milliseconds. Those two figures describe different parts of the request lifecycle — the time-to-first-token number reflects how long a caller waits before streaming begins, while the tokens-per-second figure reflects how quickly the response body arrives afterward. Both vary by provider, region, prompt length, and load, so treat them as reference points rather than guarantees.
Beyond those serving metrics, we do not currently track a confirmed context window, modality list, or public benchmark suite for GPT-6 Luna in our database. Readers evaluating it for a specific workload should check OpenAI's model documentation for the authoritative capability sheet, and use the pricing table on this page to compare what individual providers charge for access.
Common Use Cases
GPT-6 Luna is positioned as a general chat model, which makes it a candidate for conversational assistants, drafting and rewriting text, summarization, question answering over supplied context, and back-office automation where responses are generated and streamed to a user. Its measured throughput of roughly 151 tokens per second means long-form responses render at a readable pace in a streaming UI, while the ~1.46 second time to first token is acceptable for chat but less suited to autocomplete, voice turn-taking, or other sub-second interaction loops. Because our database does not record a confirmed context window or multimodal support for this model, teams planning retrieval-heavy pipelines, very long document ingestion, or image input should confirm those capabilities in OpenAI's documentation before committing. For batch or high-volume classification work, compare its per-token cost in the pricing table against smaller models, since serving speed alone rarely determines total job cost.
Frequently Asked Questions
How much does GPT-6 Luna cost to use?
Pricing depends on which provider you route through and on the pricing type — input tokens, output tokens, and any cached or batch rates are billed separately, and providers change rates frequently. Check the pricing table on this page for the current per-provider figures rather than relying on a quoted number.
What is GPT-6 Luna best used for?
It is a chat model, so it fits conversational assistants, text drafting and editing, summarization, and question answering against context you supply in the prompt. Its measured ~151 tokens per second output speed suits streaming chat interfaces where the user reads the response as it arrives.
How fast is GPT-6 Luna?
Artificial Analysis measurements recorded in our database put output throughput at about 150.7 tokens per second with a time to first token around 1,460 milliseconds. Those are averages — actual numbers shift with provider, prompt length, and traffic levels.
What context window does GPT-6 Luna support?
We do not currently have a confirmed context window recorded for GPT-6 Luna in our database. Consult OpenAI's official model documentation for the authoritative limit before designing a long-context or retrieval-heavy pipeline around it.
Is GPT-6 Luna a good choice for latency-critical applications?
Its roughly 1.5 second time to first token is fine for typical chat, where a short pause before streaming begins is expected, but applications like voice turn-taking or inline code autocomplete generally target lower initial latency. Compare its time-to-first-token figure against other models listed on this site if that metric drives your decision.