Skip to content
VECTOR WIREAI INTELLIGENCE
UTC

Gemini 3.1 Pro vs GPT-5.6 Luna

52 SHARED BENCHMARKS

Across 52 shared benchmarks, Gemini 3.1 Pro scores higher on 22 and GPT-5.6 Luna on 30. The widest gap is DeepSWE, where GPT-5.6 Luna scores 67.2 against 10. GPT-5.6 Luna is the cheaper of the two on tracked API pricing ($0.20 against $2.00 per million input tokens).

GOOGLEVSOPENAI52 SHARED22–30 HEAD-TO-HEADNEWEST SCORE ADDED

At a glance

MakerGoogleOpenAI
Released
Price per 1M tokens input / output$2.00 / $12.00$0.20 / $1.20
Cost of 1M in + 1M out$14.00$1.40 10.0× less
Head-to-head of 52 shared benchmarks22 wins30 wins
Scores tracked independently verified215 30 ◆201 43 ◆

Release dates per Artificial Analysis. Prices: Google's own price page for Gemini 3.1 Pro; OpenAI's own price page for GPT-5.6 Luna. ◆ = independently verified score.

Where each leadsby capability area · benchmark wins

AreaGemini 3.1 ProWINSGPT-5.6 Luna
Reasoning52Gemini 3.1 Pro leads 5 of 7 · widest: SimpleBench 79.6 vs 46.8
Coding16GPT-5.6 Luna leads 6 of 7 · widest: SWE-bench Pro 54.2 vs 62.7
Agentic25GPT-5.6 Luna leads 5 of 7 · widest: GDPVal 13.8 vs 47.2
Factuality30Gemini 3.1 Pro leads 3 of 3 · widest: AA-Omniscience · Non-hallucination 49.1 vs 7.4
Instruction Following10Gemini 3.1 Pro leads 1 of 1 · widest: LiveBench · Instruction Following 79.1 vs 60.1
Long Context02GPT-5.6 Luna leads 2 of 2 · widest: MRCR v2 8 needle 128k (average) 26.3 vs 74.8
Math23GPT-5.6 Luna leads 3 of 5 · widest: FrontierMath Tier 4 26.8 vs 58.5
Multimodal30Gemini 3.1 Pro leads 3 of 3 · widest: MMMU-Pro 82.4 vs 78.6

Biggest gaps

Gemini 3.1 Pro pulls furthest ahead on

  1. AA-Omniscience · Non-hallucination49.1 vs 7.4
  2. SimpleQA Verified73.5 vs 41
  3. SimpleBench79.6 vs 46.8

GPT-5.6 Luna pulls furthest ahead on

  1. GDPVal47.2 vs 13.8
  2. Terminal-Bench 4.011.6 vs 4
  3. MRCR v2 8 needle 128k (average)74.8 vs 26.3

Every shared benchmark52 · grouped by area

Reasoning 7

ARC-AGI-30.4 ◆0.2
ARC-AGI-277.1◆ 59.6
CritPt17.720.6
GPQA Diamond94.191.1
SimpleBench79.6 ◆◆ 46.8

Coding 7

Agentic 7

BrowseComp85.983.3
GDPVal13.847.2
Terminal-Bench 4.0411.6

Instruction Following 1

Long Context 2

Math 5

AIME 202698.3 ◆97.6
HMMT Feb. 202694.7 ◆98.5

Multimodal 3

LMArena · Vision1296 ◆◆ 1259
CharXiv (RQ)83.382.7
MMMU-Pro82.478.6

Other shared benchmarks 17

Each score is the one the model's own page shows — the most authoritative tracked result for that benchmark, on the benchmark's standard methodology (source-attributed). ↓ marks lower-is-better metrics; ◆ an independently verified score. Open either model for its full surface, provenance and pricing. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (google-official, openai-official), otherwise the lowest tracked offer.

Compare Gemini 3.1 Pro withALL PAIRINGS →

Compare GPT-5.6 Luna withALL PAIRINGS →