Skip to content
SOLVENCY

Head to head

GPT-5.6 Luna vs Gemini 3.1 Pro (preview)

GPT-5.6 Luna: currentGemini 3.1 Pro (preview): currentheavy tier

No verdict

Cost per solved task needs a published pass rate for both models. At least one has none, so it is reported as missing rather than estimated.

verified 2026-08-21 · prices from each provider's pricing page

Input $/M tokens

lower is better

  • GPT-5.6 Luna$0.2
  • Gemini 3.1 Pro (preview)$2

Output $/M tokens

lower is better

  • GPT-5.6 Luna$1.2
  • Gemini 3.1 Pro (preview)$12

Cost per solved task

needs a pass rate for both

  • GPT-5.6 Lunamissing
  • Gemini 3.1 Pro (preview)$4.30
GPT-5.6 LunaGemini 3.1 Pro (preview)
Input $/Mtok$0.2$2
Output $/Mtok$1.2$12
Cached input $/Mtok$0.02$0.2
Pass ratemissing46%
Pass rate sourcemissingSEAL
Cost basismissingmodelled
Cost per solved taskmissing$4.30

verified 2026-08-21 · heavy task tier; measured rows ignore the tier· Methodology· GPT-5.6 Luna· Gemini 3.1 Pro (preview)