Model results · OpenAI

GPT-6 Luna

Every published VulcanBench measurement of GPT-6 Luna, newest first: combined score, Code quality, runtime, tokens and API-equivalent cost at every effort level, graded by deterministic hidden tests on real tasks. Measured through the Codex CLI on a ChatGPT subscription. GPT-6 Luna is a newer model than GPT-5.6 Luna, which has its own page.

Suitev4

GPT-6 Luna across every effort level

In Codex across five effort levels: combined score 40.83 at Low, 45.72 at Medium, 57.35 at High, 78.79 at Extra-high and 81.42 at Max over judged runs. Six runs hit the 3-hour bound (2 at Extra-high, 4 at Max); counting them as 0 gives 71.94 at Extra-high and 67.26 at Max. Tasks passed 0, 0, 2, 11 and 12 of 23; Code quality 65.84 to 71.51; 3.6 to 60.5 minutes per task.

September 28, 2026 · 23 tasks · 5 effort levels · 115 runs of GPT-6 Luna · VulcanBench Frontier v4 · Code quality protocol v3.16

→

Suitev4

Cost and tokens across effort levels

$0.010 to $0.160 per priced task at list API rates and 0.40M to 8.47M raw tokens per task across the five effort levels; the whole sweep prices at $7.26 for 109 finished runs. The six timeouts are unpriced, not $0.

September 28, 2026 · API-equivalent estimates at list rates, solver inference only · VulcanBench Frontier v4

→

How would GPT-6 Luna do on your stack?

Public results answer "which model is stronger on this suite." Whether GPT-6 Luna belongs in your routing table depends on your languages, your codebases, and your task mix, that gets measured, not guessed. Grading rules are in the methodology.

Optimize your model routing All reports