Noometry

Model comparison

Gemini 3.1 Pro Preview vs Phi-4

Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 31.2 on the Noometry Index. Phi-4 costs 51× 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 . 23 shared benchmarks.

Gemini 3.1 Pro Preview Google

56.7

Rank #23 Confirmed

Phi-4 Microsoft

31.2

Rank #279 Confirmed

Summary

  • They share 23 benchmarks with published results for both. Gemini 3.1 Pro Preview scores higher in 9 categories and Phi-4 in 0 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Gemini 3.1 Pro Preview leads 71.7 to 17.7.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 95.6% for Gemini 3.1 Pro Preview and 13.8% for Phi-4.
  • Phi-4 is cheaper at $0.07 / $0.14 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 128K.
  • Phi-4 has downloadable open weights; the other is API-only.

Side by side

Gemini 3.1 Pro PreviewPhi-4
ProviderGoogleMicrosoft
Noometry Index56.731.2
Released2026-02-192024-12-11
WeightsProprietaryOpen
Context window1.05M128K
Max output66K4K
Input $ / M tokens$2$0.07
Output $ / M tokens$12$0.14
Results tracked7137

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

Category by category

Coding Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 42.5 (#99), Phi-4: 34.4 (#239)

BenchmarkGemini 3.1 Pro PreviewPhi-4
LMArena Coding14841231
SWE-bench Verified75.6%—
DeepSWE11.7%—
LMArena WebDev1447—
SciCode58.9%—
GSO22.6%—
WeirdML72.1%—
BigCodeBench Instruct—45.5%
LiveBench Coding—30.7%
MirrorCode8.9%—
BigCodeBench Complete—55.4%
ALE-Bench1,161—
AlgoTune2.02—

Agentic & Tool Use Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 37.7 (#34), Phi-4: 22.8 (#128)

BenchmarkGemini 3.1 Pro PreviewPhi-4
BALROG57%11.6%
Terminal-Bench80.2%—
APEX-Agents35.3%—
Berkeley Function Calling Leaderboard—28.8%
τ²-bench Banking26%—
DeepResearch Bench47.8%—
PostTrainBench22%—
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 (#12), Phi-4: 17.7 (#291)

BenchmarkGemini 3.1 Pro PreviewPhi-4
Chess Puzzles55%1%
LMArena Hard Prompts14851220
Epoch Capabilities Index154.77130.42
ARC-AGI-277.1%—
SimpleBench79.6%—
NYT Connections (extended)97.4%—
ARC-AGI-198%—
CritPt17.7%—
EnigmaEval36.8%—
Thematic Generalization79.4%—
EBR-Bench14.3%—
LiveBench Reasoning—47.8%
Mystery Game Puzzles34%—
DTBench97.1%—
LiveBench Data Analysis—45.2%
LMCA53.8%—
ForecastBench59—
LiveBench—41.6%

Math Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 62.1 (#34), Phi-4: 20.8 (#285)

BenchmarkGemini 3.1 Pro PreviewPhi-4
OTIS Mock AIME 2024-202595.6%13.8%
LMArena Math14851246
FrontierMath (Tiers 1-3)59.6%—
FrontierMath Tier 426.8%—
MathArena Final-Answer Competitions86.5%—
ProofBench26%—
LiveBench Math—42%
MATH Level 5—64.9%
FrontierMath (Feb 2025 set)36.9%—
FrontierMath Tier 4 (v1)16.7%—

Knowledge Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 71.8 (#3), Phi-4: 32.6 (#209)

BenchmarkGemini 3.1 Pro PreviewPhi-4
GPQA Diamond94.4%56.1%
Vectara Hallucination Rate10.4%3.7%
LMArena Expert14851203
Humanity's Last Exam46.4%—
SimpleQA Verified73.5%—
Confabulations—29.4%
MMLU—84.8%

Multimodal Not comparable

Gemini 3.1 Pro Preview: 37.9 (#69), Phi-4: —

BenchmarkGemini 3.1 Pro PreviewPhi-4
LMArena Vision1296—
Blueprint-Bench 226.5%—
Furniture Assembly26.7%—
LMArena Document1444—

Multilingual Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 57.0 (#12), Phi-4: 37.2 (#237)

BenchmarkGemini 3.1 Pro PreviewPhi-4
LMArena Non-English14771197
LMArena Chinese15291212
LMArena French14871224
LMArena German14911222
LMArena Japanese14931158
LMArena Korean14551151
LMArena Russian14981209
LMArena Spanish14791234

Instruction Following Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 77.0 (#32), Phi-4: 60.4 (#251)

BenchmarkGemini 3.1 Pro PreviewPhi-4
LMArena Instruction Following14661201
LiveBench Instruction Following—58.4%

Long Context Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 47.4 (#18), Phi-4: 36.9 (#226)

BenchmarkGemini 3.1 Pro PreviewPhi-4
LMArena Longer Query14831217
CL-bench20.8%—
CL-bench Life16.9%—

Writing & Preference Gemini 3.1 Pro Preview leads

Gemini 3.1 Pro Preview: 66.1 (#37), Phi-4: 40.5 (#244)

BenchmarkGemini 3.1 Pro PreviewPhi-4
LMArena Text14811217
LMArena Creative Writing14821182
LMArena Multi-Turn14881206
Short-Story Creative Writing—62.6%
EQ-Bench Creative Writing1491—
EQ-Bench 41142—
LiveBench Language—25.6%

Frequently asked questions

Is Gemini 3.1 Pro Preview better than Phi-4?

Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 31.2 on the Noometry Index. Phi-4 costs 51× 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 Phi-4?

Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Gemini 3.1 Pro Preview lists at $2 and $12.

Is Gemini 3.1 Pro Preview or Phi-4 better for coding?

Gemini 3.1 Pro Preview scores higher on coding benchmarks: 42.5 versus 34.4 in the Noometry coding category.

Which has the bigger context window?

Gemini 3.1 Pro Preview does, with 1.05M tokens against 128K.

How many benchmarks do Gemini 3.1 Pro Preview and Phi-4 share?

23 benchmarks have published results for both models. Gemini 3.1 Pro Preview has 71 scored results on Noometry and Phi-4 has 37.

Related comparisons

Go deeper