GPT-5.6 Luna Official Specifications, Pricing & Benchmarks
GPT-5.6 Luna (gpt-5.6-luna) by OpenAI, released 2026-07-09, featuring 1,050,000 context window, 128,000 max output tokens, and Text+Image+PDF input.
📋 GPT-5.6 Luna Official API & System Card Specifications
Verified against official OpenAI developer documentation & pricing
1. Identity & Capabilities
API Spec- API Model ID
- gpt-5.6-luna
- Release Date
- 2026-07-09
- Weight License
- Proprietary (Undisclosed)
- Native Reasoning
- ✅ Supported
- Function Calling & JSON Schema
- ✅ 原生支持
- Batch API 异步批处理
- ✅ Supported (50% off)
2. Context, Output & Modalities
Limits- Max Context Window
- 1,050,000 tokens (1050K)
- Max Output Tokens
- 128,000 tokens
- Input Modalities
- Text + Image + PDF
- Output Modalities
- Text
- Direct China Access
- 🌐 Global / OpenRouter
- Deployment Mode
- Cloud API only (Proprietary)
3. Official Pricing & Cache
USD / 1M- Input / 1M Tokens
- $ 0.2
- Cached Input / 1M
- $ 0.02 -90%
- Output / 1M Tokens
- $ 1.2
- Tiered Pricing Notes
🎛️ Official Sampling & Reasoning Parameters
Recommended generation settings & API parameter constraints
📊 GPT-5.6 Luna Full Benchmark Scorecard
46 public benchmarks · ranks among 154 models · synced 2026-11-20
💻 Coding 17 items
Capture-the-Flag Challenges is OpenAI's internal expansion of competitive, professional-level cybersecurity capture-the-flag tasks used to evaluate vulnerability identification and exploitation capability under the Preparedness Framework.
Terminal-Bench 2.1 is an updated release of the Terminal-Bench benchmark that tests AI agents' ability to operate a computer via the terminal.
BenchCAD variant evaluated with access to a Python tool for programmatic CAD reasoning.
DeepSWE is a software engineering agent benchmark evaluated with the mini-swe-agent harness, where each task is solved in an isolated container with no internet access. It measures an agent's ability to autonomously resolve real-world coding issues end to end.
DeepSWE 1.1 evaluates software engineering agents using the mini-swe-agent harness.
BenchCAD is a benchmark for programmatic CAD reasoning built from 17,900 execution-verified CadQuery programs spanning 106 industrial part families, roughly half anchored to real ISO, DIN, EN, ASME, and IEC specification tables.
SWE-Bench Pro is an advanced version of SWE-Bench that evaluates language models on complex, real-world software engineering tasks requiring extended reasoning and multi-step problem solving.
The Internal Research Debugging Evaluation measures whether models can debug 41 real bugs from internal OpenAI research experiments (plus alignment-auditing tasks), where the original solutions took experienced researchers hours to days.
SEC-bench Pro is a self-evolving software-security benchmark that measures agent bug hunting on critical, high-complexity systems.
The RSI (Recursive Self-Improvement) Index is OpenAI's aggregate metric across a bundle of internal AI-research evaluations, including debugging research systems, optimizing kernels and training recipes, and improving other models, measuring progress toward recursive…
FrontierCode 1.1 evaluates whether coding-agent changes are mergeable, using unit tests, maintainer-defined rubrics, and verifiers. Runs flagged for unfair internet use receive a zero score.
ExploitBench is a cybersecurity benchmark that evaluates a model's ability to discover and exploit software vulnerabilities, reported as the fraction of challenges where the model captures the target (Cap%).
PostTrainBench Lite measures whether an agent can design and execute a full post-training strategy (data, prompts, RL recipe, and eval loop) for a pretrained base model under a constrained time budget, scored as normalized mean reward over the improvement window.
KernelGen 1P is an OpenAI AI-self-improvement evaluation that measures whether models can write and optimize compute kernels, part of the suite tracking progress toward accelerating internal research.
Terminal-Bench 4.0 evaluates AI agents on real-world work in terminal and command-line containerized environments across 66 tasks, with emphasis on science-adjacent and frontier engineering problems.
ExploitGym is a large-scale, realistic benchmark built from real-world vulnerabilities across userspace programs, Google's V8 engine, and the Linux kernel.
NanoGPT is an OpenAI AI-self-improvement evaluation that measures whether models can optimize training recipes for small GPT-style models, part of the suite tracking progress toward accelerating internal research.
🤖 Agents & Tools 11 items
Connectors is an OpenAI internal production benchmark measuring reliable use of connector-based tools in agentic workflows, reported as a pass rate.
Search and Function-Calling is an OpenAI internal production benchmark measuring reliable search-tool use and function calling in agentic workflows, reported as a pass rate.
BrowseComp is a benchmark comprising 1,266 questions that challenge AI agents to persistently navigate the internet in search of hard-to-find, entangled information.
Tool Decathlon is a comprehensive benchmark for evaluating AI agents' ability to use multiple tools across diverse task categories. It measures proficiency in tool selection, sequencing, and execution across ten different tool-use scenarios.
LifeSciBench is an expert-authored, expert-reviewed benchmark of 750 open-ended life-science research tasks spanning seven workflows and seven biological domains.
Agents' Last Exam is a challenging benchmark for AI agents on hard, long-horizon tasks that test sustained reasoning, planning, and tool use, reported with and without tool access.
OSWorld 2.0 is a benchmark of 108 long-horizon, real-world computer-use workflows spanning everyday and professional tasks.
Big Finance Bench evaluates models on complex financial-analysis tasks that require retrieving and reasoning over financial documents and performing multi-step quantitative work.
Management Consulting Tasks is an internal OpenAI evaluation of long-horizon professional knowledge work drawn from management-consulting workflows, scoring whether models produce correct, decision-ready analyses.
AutomationBench is a tool-use benchmark that evaluates AI agents on automating real-world workflows, testing their ability to orchestrate tools and complete multi-step automation tasks.
GeneBench-Pro is a research-level benchmark of 129 multi-stage computational-biology problems spanning genomics, quantitative biology, and translational biomedicine.
📐 Math 2 items
A benchmark of hundreds of original, exceptionally challenging mathematics problems crafted and vetted by expert mathematicians, covering most major branches of modern mathematics from number theory and real analysis to algebraic geometry and category theory.
FrontierMath Tier 4 subset from the v2 evaluation release.
👁️ Multimodal 3 items
MMMU-Pro variant evaluated with tool access enabled.
A more robust multi-discipline multimodal understanding benchmark that enhances MMMU through a three-step process: filtering text-only answerable questions, augmenting candidate options, and introducing vision-only input settings.
GDP.pdf is a knowledge-work vision benchmark that evaluates models on economically valuable professional tasks presented as visual documents (PDFs), testing document-based reasoning, chart and table interpretation, and problem solving without tools.
📜 Long Context 5 items
A graph reasoning benchmark that evaluates language models' ability to perform breadth-first search (BFS) operations on graphs with context length over 128k tokens, testing long-context reasoning capabilities.
GraphWalks BFS variant evaluated on 1M-token contexts.
MRCR v2 (8-needle) is a variant of the Multi-Round Coreference Resolution benchmark that includes 8 needle items to retrieve from long contexts. This tests models' ability to simultaneously track and reason about multiple pieces of information across extended conversations.
MRCR v2 8-needle variant evaluated on contexts from 512K to 1M tokens.
MRCR v2 (Multi-Round Coreference Resolution version 2) is an enhanced version of the synthetic long-context reasoning task.
🧠 Reasoning & Science 3 items
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
MedChemBench is an internal OpenAI evaluation of medicinal-chemistry reasoning, testing whether models can support drug-discovery-relevant chemistry analysis and decision-making.
ARC-AGI-3 is the third-generation Abstraction and Reasoning Corpus benchmark, an interactive-reasoning evaluation designed to measure fluid, novel problem-solving ability that remains far from saturated for frontier models.
🏥 Domains 4 items
HealthBench Consensus is a HealthBench subset focused on questions where physician-created rubric criteria have especially high agreement, measuring healthcare performance and safety on consensus-evaluable conversations.
An open-source benchmark for measuring performance and safety of large language models in healthcare, consisting of 5,000 multi-turn conversations evaluated by 262 physicians using 48,562 unique rubric criteria across health contexts and behavioral dimensions.
HealthBench Professional evaluates model capability and safety for clinician use cases using real clinician-style chats and physician-authored grading rubrics.
A challenging variation of HealthBench that evaluates large language models' performance and safety in healthcare through 5,000 multi-turn conversations with particularly rigorous evaluation criteria validated by 262 physicians from 60 countries.
Source: ZeroEval / llm-stats.com (✓ = third-party verified, others vendor-reported).
Closest Self-Hostable Open-Weight Alternatives to GPT-5.6 Luna
Open-weight models with full HuggingFace weights for private GPU / Mac deployment
Frequently Asked Questions
Is GPT-5.6 Luna open-source or proprietary?
GPT-5.6 Luna (gpt-5.6-luna) is a proprietary commercial model by OpenAI. Weights and internal layer dimensions are undisclosed; it is accessed via cloud API.
What are the context limit and API pricing of GPT-5.6 Luna?
GPT-5.6 Luna supports 1050K input context and 128,000 max output tokens, priced at $0.2/1M input ($0.02/1M cached) and $1.2/1M output.