GPT-5.6 Luna (max)
OpenAIGPTProprietary
Description
GPT-5.6 Luna is the fastest and most cost-efficient tier of OpenAI's GPT-5.6 family, designed for cost-sensitive, high-volume workloads. It brings strong capability at OpenAI's lowest GPT-5.6 price, supports max reasoning effort, and has a 1.05M-token context window.
Release Date
2026-07-09
Parameters
—
Context Length
1.1M
Modalities
image, pdf, text
Capability Radar
50
general
68
coding
91
reasoning
67
science
70
agents
85
multimodal
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Agentic Capability | 50 | 52.0 | LS |
| Code Ranking | 19 | 92.0 | AA |
| General Ranking | 32 | 83.0 | AA |
| Multimodal Ranking | 70 | 43.0 | LS |
| Science | 35 | 85.0 | AA |
Benchmark Scores (LLM Stats)
(LLM Stats (zeroeval))Agents
Connectors
99.9%SR
Search and Function-Calling
89.7%SR
Capture-the-Flag Challenges (Internal)
85.2%SR
Toolathlon
53.4%SR
LifeSciBench
51.2%SR
Internal Research Debugging Evaluation
50.8%SR
Agents' Last Exam
50.3%SR
SEC-bench Pro
48.9%SR
RSI Index
41.9%SR
Big Finance Bench
36.0%SR
Management Consulting Tasks (Internal)
35.4%SR
PostTrainBench Lite
29.6%SR
KernelGen 1P
22.4%SR
AutomationBench
14.9%SR
ExploitGym
12.4%SR
GeneBench-Pro
10.8%SR
NanoGPT
1.7%SR
Code
BenchCAD (with Python tool)
73.9%SR
DeepSWE
67.2%SR
DeepSWE 1.1
67.0%
BenchCAD
63.1%SR
ExploitBench
33.2%SR
General
Artificial Analysis
51.0%
GDP.pdf
22.7%SR
Healthcare
HealthBench Consensus
95.1%SR
HealthBench
55.8%SR
HealthBench Professional
55.7%SR
HealthBench Hard
32.0%SR
Long Context
MRCR v2 (8-needle, 512K-1M)
41.3%SR
MRCR v2 (8-needle)
41.3%SR
Math
FrontierMath
78.6%SR
FrontierMath Tier 4 (v2)
58.5%SR
Multimodal
OSWorld 2.0
45.6%SR
Reasoning
GPQANYU + Cohere + Anthropic (2023)
92.3%SR
Terminal-Bench 2.1
84.7%SR
BrowseCompOpenAI (2025)
83.3%SR
Graphwalks BFS >128k
81.3%SR
SWE-Bench ProPrinceton NLP (2024)
62.7%SR
Graphwalks BFS 1M
51.2%SR
FrontierCode 1.1
39.8%
MedChemBench (Internal)
30.4%SR
ARC-AGI-3
0.2%SR
Vision
MMMU-Pro (with tools)
79.5%SR
MMMU-Pro
78.4%SR
AA Evaluation Indices
(Artificial Analysis)Coding Index(Artificial Analysis)71.4
Intelligence Index(Artificial Analysis)52.3
Gpqa(NYU + Cohere + Anthropic (2023))0.9
Terminalbench V2 10.8
Lcr(Artificial Analysis)0.8
Scicode(UIUC + Argonne National Lab (2024))0.5
Hle(Center for AI Safety + Scale AI (2025))0.4
Tau Banking0.3
LLM Stats Category Scores
(LLM Stats (zeroeval))Physics90
Biology90
Chemistry90
Search80
Math70
Spatial Reasoning70
Tool Calling70
Multimodal60
Vision60
Long Context50
Reasoning50
General50
Healthcare50
Agents50
Code50
Safety40
Finance40
Science30
Systems20
Pricing
Input Price$0.2 / 1M tokens
Output Price$1.2 / 1M tokens
Blended Price (3:1)$0.45 / 1M tokens
Cache Read Price$0.02 / 1M tokens
Cache Write Price$0.25 / 1M tokens
Speed
Tokens/sec138.1
Time to First Token82.02s
Time to Answer82.02s
Provider Price Ranking
Provider Price Ranking
2 providers
Cheapest: OpenAIMost Expensive: Neon
ProviderInputOutput
1OpenAICheapest
$0
$0
2Neon
$1
$6
Compare pricing across different API providers for this model.