Drafted with AI tools and edited by a human. The figures and conclusions were checked by a GPUniq editor.
TL;DR
GPT-6 Luna offers a 1,050,000-token context window and costs $0.10 per 1M tokens for standard short-context input, $0.75 per 1M tokens for long-context output. It scores 66.6 percent on DeepSWE v1.1, 37 on the Artificial Analysis Intelligence Index, and 59.3 percent on ARC-AGI-2. Available via OpenAI API, it's built for high-volume, cost-sensitive tasks needing long context and solid reasoning.
GPT-6 Luna Pricing
OpenAI's official API pricing for GPT-6 Luna:
- Standard short-context input: $0.10 per 1M tokens
- Standard short-context output: $0.50 per 1M tokens
- Cached input (short-context): $0.01 per 1M tokens
- Cache write (short-context): $0.125 per 1M tokens
- Long-context input: $0.20 per 1M tokens
- Long-context output: $0.75 per 1M tokens
These rates are lower than GPT-5.6 Luna's, especially for long-context operations.
Context Window Size
GPT-6 Luna handles up to 1,050,000 tokens in context. Maximum output per response is 128,000 tokens.
That lets you process massive documents, codebases, or extended conversations without losing track.
Benchmarks
Published scores:
-
DeepSWE v1.1 (Software Engineering): 66.6 percent Source
-
Artificial Analysis Intelligence Index (max reasoning effort): 37 Source
-
ARC-AGI-2 Abstraction & Reasoning: 59.3 percent Source
Features
- Context window: 1,050,000 tokens (128,000 max output)
- Multilingual text and image input, text output
- Vision (image) understanding
- Function calling supported only if reasoning_effort is "none"
- Efficient for summarization, extraction, classification
Limitations:
- Lower performance for complex coding and agentic workflows than GPT-6 Sol or Astra
- Function calling restrictions
- EU data residency only with Standard processing
Availability
GPT-6 Luna is available in the OpenAI API under model ID gpt-6-luna via Responses and Chat Completions.
Plans: ChatGPT Work, Codex for Plus, Pro, Business, Enterprise, and Edu.
Free and Go users can access Luna in the desktop app.
Pricing Comparison
| Model | Short Input ($/1M) | Short Output ($/1M) | Long Input ($/1M) | Long Output ($/1M) |
|---|---|---|---|---|
| GPT-6 Luna | 0.10 | 0.50 | 0.20 | 0.75 |
| GPT-5.6 Luna | - | - | - | - |
GPT-6 Luna offers lower rates, especially for long-context usage.
Suitable Use Cases
Pick GPT-6 Luna for:
- Large document or codebase processing
- Software engineering tasks needing extended context
- Reasoning and abstraction-heavy workloads
- High-volume jobs: summarization, extraction, classification
- Applications needing to retain large conversational or document context
For advanced coding or agentic workflows, consider GPT-6 Sol or Astra.
Related on the GPUniq blog
Methodology
Facts are compiled from the vendor's own announcement and documentation, linked below, at the time of writing; where the vendor hasn't published a number it is marked as such. GPUniq prices, when the model is already available here, come from our live catalog.
Sources
- 1.Models | OpenAI API — GPT-6 Luna specs — OpenAI (accessed )
- 2.Changelog | OpenAI API — Release of GPT-6 Sol & Luna — OpenAI (accessed )
- 3.GPT-6 Luna Benchmarks: Scores, Speed, and Cost Explained — Emergent.sh / Artificial Analysis (accessed )
Want cheap GPUs for your next project?
Browse live GPU prices and rent the right card in seconds — H100, A100, RTX 4090, and 50+ more models.
