NOUMENOpen app
NOUMEN

OpenAI · Efficient

GPT 5.6 Luna

Added Jul 9, 2026

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

GPT-5.6 Luna is the fast, cost-efficient model in OpenAI's GPT-5.6 series. It is suited for high-volume chat, classification, lightweight agentic workflows, and latency-sensitive reasoning tasks.

  • Reasoning
  • Tools
  • Vision
  • Structured output
  • PDF input
Input
Text, Images
Output
Text
Reasoning effort
none · low · medium · high · xhigh · max
Discovery tier
Efficient

Pricing

USD per 1M tokens
Input
$0.20
Output
$1.20
Cache read
$0.02
Cache write
$0.25

Reference prices as published by the model provider. API example

Specifications

Context window
1.1M
Max output
128K
Average output (7 days)
281 tokens
Longer than 29% of 465 models

Side by side

GPT 5.6 Luna compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

GPT 5.6 LunaOpenAIThis modelDeepSeek V4 Pro 0813DeepSeekRemove
ProviderOpenAIDeepSeek
Discovery tierEfficientOpen weights
AddedJul 9, 2026Aug 12, 2026
Limits
Context window1.05MBest1.05M
Max output128K384KBest
Average output281 tokens788 tokens
Pricing, USD per 1M tokens
Input$0.20Best$0.45
Output$1.20$0.80Best
Cache read$0.02Best$0.065
Cache write$0.25—
Performance
Throughput297.5 tok/sBest86.4 tok/s
Time to first token4.6 s2.4 sBest
Request success (24h)99.9%Best99.7%
Benchmark indexes
Intelligence Index37.3Best36.0
Coding Index71.4Best68.8
Capabilities
ReasoningYesYes
ToolsYesYes
VisionYesNo
Structured outputYesYes
PDF inputYesNo
Reasoning effortnone, low, medium, high, xhigh, maxnone, low, high, max

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • GPT 5.6 Luna
  • DeepSeek V4 Pro 0813

Intelligence Index

  • GPT 5.6 Luna37.3
  • DeepSeek V4 Pro 081336.0

GPT 5.6 Luna tested at max reasoning effort

GPT 5.6 Luna: Better than 81% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • GPT 5.6 Luna: 71.4Top
    • DeepSeek V4 Pro 0813: 68.8

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • GPT 5.6 Luna: 50.2%
    • DeepSeek V4 Pro 0813: 56.7%Top

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • GPT 5.6 Luna: 18.6%
    • DeepSeek V4 Pro 0813: 25.9%Top

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • GPT 5.6 Luna: 1,339 EloTop
    • DeepSeek V4 Pro 0813: 1,256 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • GPT 5.6 Luna: 1,460 EloTop
    • DeepSeek V4 Pro 0813: 1,449 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • GPT 5.6 Luna: 24.0%Top
    • DeepSeek V4 Pro 0813: 11.4%

  • AA-LCRv1.1

    Reasoning across long documents

    • GPT 5.6 Luna: 83.7%Top
    • DeepSeek V4 Pro 0813: 80.3%

  • MLCR-AA

    Long-context reasoning over medical records

    • GPT 5.6 Luna: 20.0%
    • DeepSeek V4 Pro 0813: —

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • GPT 5.6 Luna: 39.5%
    • DeepSeek V4 Pro 0813: 41.0%Top

  • GPQA Diamond

    Graduate-level science questions

    • GPT 5.6 Luna: 91.1%
    • DeepSeek V4 Pro 0813: 92.8%Top

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • GPT 5.6 Luna: 11.6%
    • DeepSeek V4 Pro 0813: 14.1%Top

  • SciCode

    Scientific computing in Python

    • GPT 5.6 Luna: 53.6%Top
    • DeepSeek V4 Pro 0813: 51.0%

Legacy evaluations

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • GPT 5.6 Luna: 83.7%Top
    • DeepSeek V4 Pro 0813: 80.3%

Scores from Artificial Analysis, updated Oct 26, 2026. GPT 5.6 Luna was tested at max reasoning effort; compare scores measured at the same effort.Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
297.5 tok/s
Faster than 93% of 247 models
Time to first token
4.6 s
Quicker than 35% of 247 models
Prompt cache hits
85.5%
About $0.0462 per 1M input tokens with caching
Request success (24h)
99.9%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 99.7%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-5.6-luna",
    "reasoning_effort": "medium",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="openai/gpt-5.6-luna",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

OpenAI

Other OpenAI models, starting with the same family.