Back

OpenAI

GPT-6 Luna

ReasoningClosed weightsModels.dev
Intelligence
38.1
Rank 79 of 744 measured models
Output speed
116 t/s
Rank 137 of 403 measured models
Price/1M
$0.20
Rank 161 of 659 measured models
Context window
1M
Rank 84 of 1199 measured models
Cost per task
$0.0678
Rank 42 of 193 measured models

GPT-6 Luna (Max) scores 38.1 on the Artificial Analysis Intelligence Index, ranking 79th of 744 measured models. The closest models on this index are GPT-5.5 (38.4), Claude Sonnet 5 (38.2), and MiMo-V2.6-Flash (37.9). At $0.20 per million tokens (3:1 input/output blended price), it is cheaper than GPT-5.5 ($11.25) and Claude Sonnet 5 ($4), more expensive than MiMo-V2.6-Flash ($0.175). With an output speed of 116 t/s, it is faster than GPT-5.5 (81.9 t/s), Claude Sonnet 5 (74.9 t/s), and MiMo-V2.6-Flash (53.8 t/s).

Artificial Analysis Indices
Maximum
Intelligence38.1
Rank 79 of 744 measured models
Reasoning levels
Each point is a reasoning level of this model: more reasoning often raises the score but costs more. Select a point to open that level.
Intelligence Indexmost efficient ↗Cost per task
Capability indexes
Performance in specific industries, measured by Artificial Analysis
Finance & Accounting40.3
Rank 72 of 218 measured models
Strategy & Ops45.7
Rank 61 of 218 measured models
Legal40.3
Rank 81 of 218 measured models
Healthcare & Medical37.4
Rank 39 of 82 measured models
Engineering37.1
Rank 81 of 217 measured models
Economics45.2
Rank 81 of 225 measured models
Cyber Index
Artificial Analysis · v1
Cyber Index52.7%Rank 4 of 20 measured models
Successes
52.7%
Safety blocks
0.0%
Failures
47.3%
CWE-Bench-AA56.7%
DeepsecBench-AA23.5%
CyberGym-E2E-AA77.9%
Cost per task$0.117
Cyber Index leaderboard
Decision Index
GPT-6 Luna (Max) · Artificial Analysis
Decision Index60.8Rank 9 of 26 measured models
AA-Omniscience-Decision · accuracy73.6
AA-Omniscience-Decision · non-hallucination3.0
HLE-MultiChoice35.3
When2Call79.0
WildGuardTest90.5
Cost per 1M tasks$591.63
Average time per task2.25 s
Decision Index leaderboard
Standard Benchmarks
HLE38.5%
Rank 96 of 664 measured models
SciCode54.6%
Rank 58 of 231 measured models
AA-LCR83.3%
Rank 20 of 575 measured models
Terminal-Bench v412.6%
Rank 77 of 218 measured models
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.

Privacy

Allow navigation analytics?

With your permission, Rybbit measures visits, performance and interactions (clicks, copies, forms), including session replay, to improve BenchSift. Input values are masked in replays. Refusing does not limit the site.

Your choice is stored for 180 days. Change it at any time using “Privacy preferences” in the footer. Withdrawal reloads the page.

No choice stored: Rybbit is disabled.

Privacy