Noometry

Model comparison

DeepSeek V4 Flash vs Gemini 3.1 Pro Preview

Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 53.6 on the Noometry Index. DeepSeek V4 Flash costs 17× 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 . 39 shared benchmarks.

DeepSeek V4 Flash DeepSeek

53.6

Rank #32 Confirmed

Gemini 3.1 Pro Preview Google

56.7

Rank #20 Confirmed

Summary

  • They share 39 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 1 category and Gemini 3.1 Pro Preview in 7 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Gemini 3.1 Pro Preview leads 71.7 to 53.7.
  • The biggest single-benchmark swing is SimpleQA Verified: 33.6% for DeepSeek V4 Flash and 73.5% for Gemini 3.1 Pro Preview.
  • DeepSeek V4 Flash is cheaper at $0.15 / $0.60 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 1M.
  • DeepSeek V4 Flash has downloadable open weights; the other is API-only.

Side by side

DeepSeek V4 FlashGemini 3.1 Pro Preview
ProviderDeepSeekGoogle
Noometry Index53.656.7
Released2026-04-242026-02-19
WeightsOpenProprietary
Context window1M1.05M
Max output393K66K
Input $ / M tokens$0.15$2
Output $ / M tokens$0.60$12
Results tracked4171

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

Category by category

Coding DeepSeek V4 Flash leads

DeepSeek V4 Flash: 47.6 (#56), Gemini 3.1 Pro Preview: 42.7 (#88)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
LMArena WebDev15811446
SciCode49.9%58.9%
WeirdML63%72.1%
LMArena Coding14571483
ALE-Bench1,3061,161
SWE-bench Verified—75.6%
DeepSWE—11.7%
FrontierCode18.8%—
GSO—22.6%
MirrorCode—8.9%
AlgoTune—2.02

Agentic & Tool Use Not comparable

DeepSeek V4 Flash: —, Gemini 3.1 Pro Preview: 37.7 (#33)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
Terminal-Bench—80.2%
APEX-Agents—35.3%
τ²-bench Banking—26%
DeepResearch Bench—47.8%
PostTrainBench—22%
BALROG—57%
ExploitBench—26.1%
GBAEval—0.8%
GDP.pdf—17%
LMArena Search—1211
METR Time Horizons—77%
Vending-Bench 2—3,774

Reasoning Gemini 3.1 Pro Preview leads

DeepSeek V4 Flash: 53.7 (#25), Gemini 3.1 Pro Preview: 71.7 (#9)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
ARC-AGI-261.4%77.1%
SimpleBench61.1%79.6%
NYT Connections (extended)89.6%97.4%
ARC-AGI-189%98%
CritPt16.6%17.7%
Chess Puzzles33%55%
LMArena Hard Prompts14431484
Mystery Game Puzzles34%34%
DTBench90.9%97.1%
LMCA41.7%53.8%
Epoch Capabilities Index154.51154.92
Kagi LLM Benchmark52.2%—
EnigmaEval—36.8%
Thematic Generalization—79.4%
EBR-Bench—14.3%
ForecastBench—59

Math Gemini 3.1 Pro Preview leads

DeepSeek V4 Flash: 60.0 (#31), Gemini 3.1 Pro Preview: 61.8 (#28)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
FrontierMath (Tiers 1-3)57.5%59.6%
FrontierMath Tier 424.4%26.8%
MathArena Final-Answer Competitions76.5%86.5%
OTIS Mock AIME 2024-202594.4%95.6%
ProofBench56%26%
LMArena Math14271485
FrontierMath (Feb 2025 set)—36.9%
FrontierMath Tier 4 (v1)—16.7%

Knowledge Gemini 3.1 Pro Preview leads

DeepSeek V4 Flash: 55.5 (#43), Gemini 3.1 Pro Preview: 71.8 (#3)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
GPQA Diamond91%94.4%
SimpleQA Verified33.6%73.5%
LMArena Expert14411484
Humanity's Last Exam—46.4%
Vectara Hallucination Rate—10.4%

Multimodal Not comparable

DeepSeek V4 Flash: —, Gemini 3.1 Pro Preview: 38.1 (#60)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
LMArena Vision—1296
Blueprint-Bench 2—26.5%
Furniture Assembly—26.7%
LMArena Document—1444

Multilingual Gemini 3.1 Pro Preview leads

DeepSeek V4 Flash: 53.2 (#71), Gemini 3.1 Pro Preview: 57.1 (#11)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
LMArena Non-English14201475
LMArena Chinese14671528
LMArena French14381488
LMArena German14181491
LMArena Japanese14031494
LMArena Korean13831458
LMArena Russian14281496
LMArena Spanish14371480

Instruction Following Gemini 3.1 Pro Preview leads

DeepSeek V4 Flash: 75.1 (#76), Gemini 3.1 Pro Preview: 77.1 (#31)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
LMArena Instruction Following14211465

Long Context Gemini 3.1 Pro Preview leads

DeepSeek V4 Flash: 43.8 (#81), Gemini 3.1 Pro Preview: 47.3 (#17)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
LMArena Longer Query14351482
CL-bench—20.8%
CL-bench Life—16.9%

Writing & Preference Gemini 3.1 Pro Preview leads

DeepSeek V4 Flash: 64.0 (#59), Gemini 3.1 Pro Preview: 66.3 (#36)

BenchmarkDeepSeek V4 FlashGemini 3.1 Pro Preview
LMArena Text14321480
LMArena Creative Writing14031482
EQ-Bench Creative Writing15591491
LMArena Multi-Turn14491486
EQ-Bench 4—1142

Frequently asked questions

Is DeepSeek V4 Flash better than Gemini 3.1 Pro Preview?

Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 53.6 on the Noometry Index. DeepSeek V4 Flash costs 17× 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, DeepSeek V4 Flash or Gemini 3.1 Pro Preview?

DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Gemini 3.1 Pro Preview lists at $2 and $12.

Is DeepSeek V4 Flash or Gemini 3.1 Pro Preview better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.6 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 1M.

How many benchmarks do DeepSeek V4 Flash and Gemini 3.1 Pro Preview share?

39 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and Gemini 3.1 Pro Preview has 71.

Related comparisons

Go deeper