GPT-6 Luna
GPT-6 Luna is OpenAI's most efficient model for cost-sensitive, high-volume workloads, with 1.05M context, adjustable reasoning and full tool support.
Model Overview
Capabilities, design details, and architectural traits
GPT-6 Luna - OpenAI's most efficient model for focused, high-volume tasks
GPT-6 Luna is the efficiency tier of OpenAI's GPT-6 flagship lineup. Where GPT-6 Astra targets the hardest end-to-end work and GPT-6 Sol balances intelligence and cost, Luna is built for cost-sensitive, high-volume workloads where efficiency matters more than maximum reasoning depth.
| Trait | Detail |
|---|---|
| Positioning in the GPT-6 lineup | The most efficient model, recommended for focused, high-volume tasks instead of complex reasoning and coding |
| Reasoning effort | Adjustable from none up to max, so it can skip reasoning entirely for simple, high-volume requests |
| Knowledge cutoff | The most recent cutoff among the GPT-6 flagship models |
| Voice-agent role | Paired with GPT-Live-1 as the backend model for high-volume tasks like scheduling and order updates |
Where Luna fits in a workload
Luna shares the flagship feature set: a 1.05M token context window, 128K max output, text and image input, vision, multilingual capabilities, and the same tool support as its siblings, including Functions, Web search, File search, and Computer use. The difference is efficiency. OpenAI positions it as the model to reach for when a task does not need Astra-level reasoning, such as routine agent steps or bulk processing.
Splitting work across models
The GPT-Live-1 API release shows the intended pattern: developers pair the voice model with Luna for high-volume tasks like scheduling or order updates, and switch to Astra for complex customer issues that require deeper reasoning. Luna is the cost-efficient workhorse in that split, matching reasoning depth and cost to each task.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | OpenAI | Anthropic | Anthropic |
| Release Date | September 22, 2026 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | May 2026 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 1.1M Best Context Window | 1M | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.10 Best Input Pricing | $4 | $2 |
| Output Pricing | $0.50 Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | textimagefile | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 38.1 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | - | - | - |
| Agentic Index | - | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from OpenAI
Other models by OpenAI
Top AI Models
Leading alternatives by intelligence score
Claude Opus 5.5
Muse Spark 1.3
Gemini 4 Argon
