LLM Reference

GPT-5.6 Luna

Released
2026-07-09
Last refreshed
2026-07-09
Status
Researched 57d ago
ProprietaryCommercial use: conditionalMultimodalRAGAgentsLong contextVisionJSON / Tool use

GPT-5.6 Luna is worth evaluating for rag, agents, and long context when its provider route and context window match the workload.

Use it for

  • Teams evaluating rag, agents, and long context
  • Workloads that can use a 1.05m context window
  • Buyers comparing 2 tracked provider routes

Do not use it for

  • Workloads where another current model has stronger sourced task evidence
Specifications
Family
GPT-5.6
Released
2026-07-09
Context
1.05m
Max output
128,000
Architecture
Decoder Only
Specialization
general
Openness
Proprietary
License
ProprietaryCommercial use: conditional
Weights
Not released
Code
Unknown
Created by

Cutting-edge research and development.

San Francisco, California, United States
Founded 2015
Website
Pricing
Output / 1M
$6.00
Input / 1M
$1.00

Cheapest of 2 routes · OpenAI API · cache read $0.100

About

OpenAI's fast, low-cost GPT-5.6 model. Luna is the affordable tier in the Sol, Terra, and Luna lineup, optimized for latency-sensitive applications such as summarization, drafting, autocomplete, and routine automation. Generally available July 9, 2026 across ChatGPT, Codex, and the OpenAI API (model ID gpt-5.6-luna). Supports text and image inputs, reasoning, tool use, prompt caching, and Batch API with a 1,050,000-token context window and 128K max output.

Top use-case fit: coding, agents, and build tasks

RAG

Included by capability and metadata signals in the decision map.

Agents

Included by capability and metadata signals in the decision map.

Long context

Included by capability and metadata signals in the decision map.

Provider price ladder

Compare all 2

Compare API pricing across 2 providers for input and output tokens, batch, and cached reads when available.

ProviderInput / 1MOutput / 1MBatch in / outCacheRoute
OpenAI API$1.00$6.00$0.500 / $3.00read $0.100
Serverless
OpenRouter$1.00$6.00--
Serverless

Available via routers & gateways(15)

Capabilities

VisionMultimodalReasoningJSON / Tool usePrompt CachingBatch API

Benchmark peer barsfor RAG

No task-mapped benchmark peers are available for this model yet.

Benchmark scores(10)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionEvaluationSource
HealthBench Professional55.7length-adjusted; raw 59.8; mean response 3389 charsObserved 2026-06-26Source
HealthBench55.8length-adjusted; raw 55.4; mean response 1930 charsObserved 2026-06-26Source
HealthBench Hard32.0length-adjusted; raw 31.4; mean response 1923 charsObserved 2026-06-26Source
HealthBench Consensus95.1length-adjusted; raw 95.1; mean response 1897 charsObserved 2026-06-26Source
GeneBench-Pro23.6GPT Pro (Extended); max reasoning comparatorObserved 2026-06-30Source
CursorBench61.1CursorBench 3.2Observed 2026-07-18
Configuration: GPT-5.6 Luna Max
Harness: CursorBench 3.2 productized Cursor-agent workflow
Evaluator: Cursor
Confidence: confirmed
Cost/task: $1.97
Tokens/task: 87,973
Steps/task: 61
Notes: Cursor vendor-reported result; not independently reproducible. Results are subject to variance, and small score differences may not be statistically meaningful.
Source
CursorBench57.7CursorBench 3.2Observed 2026-07-18
Configuration: GPT-5.6 Luna Extra High
Harness: CursorBench 3.2 productized Cursor-agent workflow
Evaluator: Cursor
Confidence: confirmed
Cost/task: $1.14
Tokens/task: 22,480
Steps/task: 48
Notes: Cursor vendor-reported result; not independently reproducible. Results are subject to variance, and small score differences may not be statistically meaningful.
Source
CursorBench56.8CursorBench 3.2Observed 2026-07-18
Configuration: GPT-5.6 Luna High
Harness: CursorBench 3.2 productized Cursor-agent workflow
Evaluator: Cursor
Confidence: confirmed
Cost/task: $0.82
Tokens/task: 15,141
Steps/task: 40
Notes: Cursor vendor-reported result; not independently reproducible. Results are subject to variance, and small score differences may not be statistically meaningful.
Source
CursorBench47.7CursorBench 3.2Observed 2026-07-18
Configuration: GPT-5.6 Luna Medium
Harness: CursorBench 3.2 productized Cursor-agent workflow
Evaluator: Cursor
Confidence: confirmed
Cost/task: $0.39
Tokens/task: 7,095
Steps/task: 28
Notes: Cursor vendor-reported result; not independently reproducible. Results are subject to variance, and small score differences may not be statistically meaningful.
Source
CursorBench37.6CursorBench 3.2Observed 2026-07-18
Configuration: GPT-5.6 Luna Low
Harness: CursorBench 3.2 productized Cursor-agent workflow
Evaluator: Cursor
Confidence: confirmed
Cost/task: $0.16
Tokens/task: 3,209
Steps/task: 17
Notes: Cursor vendor-reported result; not independently reproducible. Results are subject to variance, and small score differences may not be statistically meaningful.
Source

Migration checks

No linked migration route is available for this model yet.