Noometry

Model comparison

Gemini 3.1 Pro Preview vs GPT-5.4 mini

Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 44.8 on the Noometry Index. GPT-5.4 mini costs 2.7× less per token, which makes it the better buy when Gemini 3.1 Pro Preview's lead doesn't matter for your workload.

Last verified . 44 shared benchmarks.

Gemini 3.1 Pro Preview Google

56.7

Rank #20 Confirmed

GPT-5.4 mini OpenAI

44.8

Rank #69 Confirmed

Summary

  • They share 44 benchmarks with published results for both. Gemini 3.1 Pro Preview scores higher in 8 categories and GPT-5.4 mini in 2 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Gemini 3.1 Pro Preview leads 71.7 to 30.3.
  • The biggest single-benchmark swing is ARC-AGI-2: 77.1% for Gemini 3.1 Pro Preview and 18.9% for GPT-5.4 mini.
  • GPT-5.4 mini is cheaper at $0.75 / $4.50 per million input/output tokens, against $2 / $12 for Gemini 3.1 Pro Preview.
  • Gemini 3.1 Pro Preview accepts more context: 1.05M tokens versus 400K.

Side by side

Gemini 3.1 Pro PreviewGPT-5.4 mini
ProviderGoogleOpenAI
Noometry Index56.744.8
Released2026-02-192026-03-17
WeightsProprietaryProprietary
Context window1.05M400K
Max output66K128K
Input $ / M tokens$2$0.75
Output $ / M tokens$12$4.50
Results tracked7146

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding GPT-5.4 mini leads

Gemini 3.1 Pro Preview: 42.7 (#88), GPT-5.4 mini: 45.0 (#67)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
LMArena WebDev14461397
SciCode58.9%49.9%
WeirdML72.1%60.3%
LMArena Coding14831438
ALE-Bench1,1611,189
SWE-bench Verified75.6%—
DeepSWE11.7%—
FrontierCode—27%
GSO22.6%—
MirrorCode8.9%—
AlgoTune2.02—

Agentic & Tool Use Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 37.7 (#33), GPT-5.4 mini: 29.6 (#74)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
DeepResearch Bench47.8%36.3%
Terminal-Bench80.2%—
APEX-Agents35.3%—
τ²-bench Banking26%—
PostTrainBench22%—
BALROG57%—
ExploitBench26.1%—
GBAEval0.8%—
GDP.pdf17%—
LMArena Search1211—
METR Time Horizons77%—
Vending-Bench 23,774—

Reasoning Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 71.7 (#9), GPT-5.4 mini: 30.3 (#77)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
ARC-AGI-277.1%18.9%
NYT Connections (extended)97.4%61.8%
ARC-AGI-198%63.7%
CritPt17.7%10%
Chess Puzzles55%24%
Thematic Generalization79.4%61.7%
LMArena Hard Prompts14841424
Mystery Game Puzzles34%11%
DTBench97.1%80%
LMCA53.8%40.8%
Epoch Capabilities Index154.92148.83
ForecastBench5957
SimpleBench79.6%—
Kagi LLM Benchmark—37.9%
EnigmaEval36.8%—
EBR-Bench14.3%—

Math Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 61.8 (#28), GPT-5.4 mini: 45.4 (#63)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
FrontierMath (Tiers 1-3)59.6%51.2%
FrontierMath Tier 426.8%9.8%
OTIS Mock AIME 2024-202595.6%88.9%
ProofBench26%21%
LMArena Math14851418
FrontierMath (Feb 2025 set)36.9%28.3%
FrontierMath Tier 4 (v1)16.7%2.1%
MathArena Final-Answer Competitions86.5%—

Knowledge Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 71.8 (#3), GPT-5.4 mini: 51.4 (#62)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
GPQA Diamond94.4%86.9%
SimpleQA Verified73.5%29.4%
Vectara Hallucination Rate10.4%5.5%
LMArena Expert14841434
Humanity's Last Exam46.4%—

Multimodal Too close to call

Gemini 3.1 Pro Preview: 38.1 (#60), GPT-5.4 mini: 39.0 (#51)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
LMArena Vision12961245
Blueprint-Bench 226.5%—
Furniture Assembly26.7%—
LMArena Document1444—

Multilingual Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 57.1 (#11), GPT-5.4 mini: 52.1 (#94)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
LMArena Non-English14751405
LMArena Chinese15281446
LMArena French14881440
LMArena German14911407
LMArena Japanese14941375
LMArena Korean14581369
LMArena Russian14961417
LMArena Spanish14801403

Instruction Following Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 77.1 (#31), GPT-5.4 mini: 74.3 (#99)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
LMArena Instruction Following14651405

Long Context Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 47.3 (#17), GPT-5.4 mini: 42.9 (#107)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
LMArena Longer Query14821407
CL-bench20.8%—
CL-bench Life16.9%—

Writing & Preference Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 66.3 (#36), GPT-5.4 mini: 64.2 (#56)

BenchmarkGemini 3.1 Pro PreviewGPT-5.4 mini
LMArena Text14801411
LMArena Creative Writing14821370
EQ-Bench Creative Writing14911665
LMArena Multi-Turn14861429
EQ-Bench 41142—

Frequently asked questions

Is Gemini 3.1 Pro Preview better than GPT-5.4 mini?

Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 44.8 on the Noometry Index. GPT-5.4 mini costs 2.7× less per token, which makes it the better buy when Gemini 3.1 Pro Preview's lead doesn't matter for your workload.

Which is cheaper, Gemini 3.1 Pro Preview or GPT-5.4 mini?

GPT-5.4 mini is cheaper. It lists at $0.75 per million input tokens and $4.50 per million output tokens; Gemini 3.1 Pro Preview lists at $2 and $12.

Is Gemini 3.1 Pro Preview or GPT-5.4 mini better for coding?

GPT-5.4 mini scores higher on coding benchmarks: 45.0 versus 42.7 in the Noometry coding category.

Which has the bigger context window?

Gemini 3.1 Pro Preview does, with 1.05M tokens against 400K.

How many benchmarks do Gemini 3.1 Pro Preview and GPT-5.4 mini share?

44 benchmarks have published results for both models. Gemini 3.1 Pro Preview has 71 scored results on Noometry and GPT-5.4 mini has 46.

Related comparisons

Go deeper