OpenAIAIWS category: Efficient Multimodal
GPT-5.6 Luna
OpenAI GPT-5.6 Luna, a GPT-5.6 model positioned for cost-sensitive, high-volume workloads.
Facts Verified
Sep 21, 2026
Verification
Verified
Profile
Source-backed
Context
1,050,000 tokens
Max output
128,000 tokens
API
Available
Specifications
Context
1,050,000 tokens
Max output
128,000 tokens
CompanyOpenAI
Release DateJul 9, 2026
Provider statusNot published
LicenseNot published
API availabilityAvailable
Capability Map
Verified fieldsReasoning
Supported
Image input
Supported
Audio input
Not supported
Video input
Not supported
API access
Available
Native input modalities
textimage
Native output: text
Built-in tools
Image generation
Supported
Web search
Supported
File search
Supported
Code interpreter
Supported
Computer use
Supported
MCP
Supported
Hosted shell
Supported
Apply patch
Supported
Skills
Supported
Tool search
Supported
Native output is what the model returns directly. Tool capabilities are separate.
Developer API Rates
Verified public rateInput / 1M
$0.2000
Output / 1M
$1.20
Cached $0.0200 / 1MAbove 272,000 tokens
Above 272K input tokens, the full request uses 2× input and 1.5× output pricing.
Cache writes are billed at 1.25× the uncached input rate.
Verified Capabilities
- 1,050,000 tokens published context window
- 128,000 tokens published maximum output
- Reasoning support listed by the provider
- Multimodal input: text, image
Pricing & Availability Notes
- Separate long-context pricing applies: Above 272,000 tokens
- Above 272K input tokens, the full request uses 2× input and 1.5× output pricing.
- Cache writes are billed at 1.25× the uncached input rate.
Provider API model ID
Use this identifier in the provider API where supported.
gpt-5.6-luna