LLM Pricing
ModelsLeaderboardsToolsProvidersReleasesArticles

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list

GPT-6 Luna

openai·openai/gpt-6-luna·GA·Closed·gpt-luna series·NEW
OpenAI's most efficient model for focused, high-volume tasks

Specs & pricing

Input / output per 1M tokens
Official price·OpenAI
$0.10 / $0.50
Blended $0.20 · Cache read $0.01
Lowest paid·KunavoGateway
$0.04 / $0.20
Blended $0.08 · 5× spread
1 more $0 channels
Context
1,050,000
Output limit
128,000
Knowledge cutoff
2026-05-18
Released / updated
2026-09-22 / 2026-09-22
Capabilities
✓ Reasoning✓ Tool use✓ Structured output✓ Temperature✓ Attachments
⚠ Providers report capability flags inconsistently
Modalities
TextImagePDF

GPT-6 Luna by openai is offered by 58 providers on this page. Public prices are shown for 56 of them. Its official list price is $0.10 per 1M input tokens and $0.50 per 1M output tokens. The lowest paid channel is Kunavo at $0.04 / $0.20 per 1M, about 2.5× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it. It also has 1 free ($0) channel; free tiers usually carry rate limits, and subscription-covered access bills $0 per token only after the subscription fee.

Artificial Analysis rates it 27.9 on the Intelligence Index, with 43.2 for Strategy & Ops, its strongest domain. Against its lowest blended price of $0.08 per 1M, that is roughly 349 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 119 tokens per second, with 101.86s to the first token. Running one task of the Artificial Analysis suite costs about $0.1520, which reflects how many tokens its reasoning consumes rather than the unit price alone.

The context window is 1,050,000 tokens at the reference host, but hosts report different limits, from 400,000 to 1,050,000, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. Accepted input modalities are Text, Image, and PDF. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use. Its training knowledge cuts off at 2026-05-18.

Quality & performanceArtificial Analysis · Intelligence Index v5 · rep. tier Max

Intelligence27.9
Value349
Engineering33.3
Finance & Accounting33.9
Legal32.2
Healthcare & Medical34.8
Strategy & Ops43.2
Economics34.5
Output speed119 tok/s
TTFT101.86 s
Cost per task$0.15
Value formulaIQ 27.9 ÷ min blended $0.080 = 349
Reasoning tier → intelligence / speed (higher tier = stronger but slower)
Non-reasoningIQ 9.4 · 110 tok/s
LowIQ 9.7 · 111 tok/s
MediumIQ 18.5 · 117 tok/s
HighIQ 21.4 · 112 tok/s
XhighIQ 23.5 · 114 tok/s
MaxIQ 27.9 · 119 tok/s

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 58 providers56 with public prices · 1 free

  • ProviderInput / Output
  • GitLab DuoCloud
    duo-chat-gpt-6-luna
    Free
    Cache read $0Context 1,050,000Output limit 128,000
  • KunavoGateway
    $0.04 / $0.20
    tiered >272K: $0.08
    Cache read $0.004Context 1,050,000Output limit 128,000
  • UplinkSwitchGateway
    $0.063 / $0.20
    Cache read $0.013Context 1,050,000Output limit 128,000
  • Outer LabsGateway
    $0.06 / $0.30
    tiered >272K: $0.12
    Cache read $0.006Context 1,050,000Output limit 128,000
  • IteraComputeGateway
    $0.08 / $0.40
    tiered >272K: $0.16
    Cache read $0.008Context 1,050,000Output limit 128,000
  • KenariGateway
    $0.10 / $0.49
    Cache read $0.01Context 1,050,000Output limit 128,000
  • OpenAIOfficialFirst-party
    $0.10 / $0.50
    tiered >272K: $0.20
    Cache read $0.01Context 1,050,000Output limit 128,000
  • Amazon BedrockCloud
    global.openai.gpt-6-luna
    $0.10 / $0.50
    tiered >272K: $0.20
    Cache read $0.01Context 1,050,000Output limit 128,000
  • Merge GatewayGateway
    $0.10 / $0.50
    tiered >272K: $0.20
    Cache read $0.01Context 1,050,000Output limit 128,000
  • SynthoraiGateway
    $0.10 / $0.50
    tiered >272K: $0.20
    Cache read $0.01Context 1,050,000Output limit 128,000
ProviderTierInputOutputCache readCache writeContextOutput limitStatus
GitLab Duo
duo-chat-gpt-6-luna
CloudFree$0$01,050,000128,000
KunavoGateway$0.04
tiered >272K: $0.08
$0.20$0.004$0.051,050,000128,000
UplinkSwitchGateway$0.063$0.20$0.013—1,050,000128,000
Outer LabsGateway$0.06
tiered >272K: $0.12
$0.30$0.006—1,050,000128,000
IteraComputeGateway$0.08
tiered >272K: $0.16
$0.40$0.008$0.101,050,000128,000
KenariGateway$0.10$0.49$0.01$0.111,050,000128,000
OpenAIOfficialFirst-party$0.10
tiered >272K: $0.20
$0.50$0.01$0.131,050,000128,000
Amazon Bedrock
global.openai.gpt-6-luna
Cloud$0.10
tiered >272K: $0.20
$0.50$0.01$0.131,050,000128,000
Merge GatewayGateway$0.10
tiered >272K: $0.20
$0.50$0.01$0.131,050,000128,000
SynthoraiGateway$0.10
tiered >272K: $0.20
$0.50$0.01—1,050,000128,000

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Reasoning control

effort = noneeffort = loweffort = mediumeffort = higheffort = xhigheffort = maxToggle (on / off)

Interleaved thinking (reasoning between tool calls) is declared by 2 of 58 providers.

Experimental modes3 items

ModeProviderInputOutputCache readCache write
fastOpenAI$0.20$1.00$0.02$0.25
flexOpenAI$0.05$0.25$0.005$0.063
proOpenAI————

Your usage cost

Common scenarios
1Kunavo$13.68
2UplinkSwitch$16.50
3Outer Labs$20.52
4IteraCompute$27.36
5Kenari$33.49
6OpenAI · Official$34.20

1 more channels offer $0 (GitLab Duo); free tiers usually have rate limits and no SLA, excluded from ranking.

Switch to Kunavo to save $20.52/mo (60%)
Note: this is a gateway; verify availability and rate limits yourself.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$0.50$02026-09-232026-11-15
2026-11-15Input list $0.10Output list $0.50Min blended $0.08

Benchmark1 items

NameConditionsScoreMetricSource
DeepSWEvariant: max effort · v1.166.6scoreSource ↗

The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.

Artificial Analysis evaluations27 items

Humanity's Last Exam38.5%
HLE-Diamond22.2%
MMMU-Pro79.7%
CritPt19.4%
CritPt v2 Hard42.5%
SciCode54.6%
FrontierCode41.9%
FrontierCode 2.025.2%
Terminal-Bench 4.012.6%
Terminal-Bench-Science8.6%
AutomationBench-AA53.2%
AA-AnalystAgent43.8%
AA-Briefcase1337
GDPval-AA1444
GDP.pdf22.8%
AA-LCR83.3%
MLCR-AA26.1%
AA-Omniscience Index-8.3
AA-Omniscience Accuracy39.6%
AA-Omniscience Non-Hallucination20.6%
Harvey LAB-AA3.3%
HealthBench54.6%
HealthBench Professional61.9%
AA-3D1000
Code2Video-AA1030
mka0.3675
vials-aa0.2733

Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and scale (percent correct, points or an index), so rows are not comparable with one another.

Related models

GPT Luna Latestsame series—GPT-5.6 Lunasame series$0.20 / $1.20Nemotron 3 Ultra 550B A55Bcheaper alternative$0 / $0Nemotron 3.5 Super VL 120B A12Bcheaper alternative$0 / $0reBeam 1.0 501B A23Bcheaper alternative$0 / $0Qwen3.7 Flashcheaper alternative$0.03 / $0.13GPT-6 Astra Prosame lab—GPT-6 Luna Prosame lab—
Data partly from models.dev (MIT) · OpenAI official docs ↗