LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

DeepSeek V4 Pro 0813 vs Gemini 3.7 Flash

Updated October 21, 2026. 72 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace DeepSeek V4 Pro 0813 Replace Gemini 3.7 Flash
DeepSeek V4 Pro 0813DeepSeek
Coding
54.6
Overall
36.0
Output $/M
$1.98
11
Gemini 3.7 FlashGoogle
Coding
69.9
Overall
39.6
Output $/M
$3.75
60
vs boards won of 72 shared, 1 tied

Decision reading

Pick Gemini 3.7 Flash for coding: it scores 69.9 on the Vibe Coding Index against 54.6 for DeepSeek V4 Pro 0813.

Pick DeepSeek V4 Pro 0813 when output tokens cost 1.9x less, or when a solved task costs $0.10 against $1.78.

Model ADeepSeek V4 Pro 0813DeepSeek54.6/1001 independent benchmark$0.10 per solved taskModel BGemini 3.7 FlashGoogle69.9/1002 independent benchmarks$1.78 per solved task
Shared results
72
DeepSeek V4 Pro 0813 only
6
Gemini 3.7 Flash only
40
Like-for-like categories
7 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
DeepSeek V4 Pro 0813 54.6 from 1 benchmark, Gemini 3.7 Flash 69.9 from 2 benchmarks.
Widest category
Coding, 9.4 points apart on average over 19 shared benchmarks.
Widest single gap
CyberBench (Vals AI): DeepSeek V4 Pro 0813 72.9% against Gemini 3.7 Flash 43.8%, 29.1 points apart on a 0 to 100 scale.
Output price
DeepSeek V4 Pro 0813 $1.98 against Gemini 3.7 Flash $3.75 per million output tokens, 1.9x apart.
Context window
Both take 1.05M tokens.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryDeepSeek V4 Pro 0813BarsGemini 3.7 FlashTies
Coding7111 tie
Intelligence02
Agentic work25
Reasoning110
Math15
Knowledge09
General018
All shared boards11601 tie
CodingLike-for-likeFour of these boards feed the Vibe Coding Index711
Coding benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
LiveBench codingNot measured80.5%Only Gemini 3.7 Flash measured
LMArena WebDev15821591Gemini 3.7 Flash
AA Coding Index68.871.5Gemini 3.7 Flash
LiveCodeBench (Vals AI)87.5%88.7%Gemini 3.7 Flash
SWE-bench Verified (Vals AI)96.4%80.8%DeepSeek V4 Pro 0813
Terminal-Bench 4 (Vals AI)14.1%12.1%DeepSeek V4 Pro 0813
Vibe Code Bench (Vals AI)82.3%70.4%DeepSeek V4 Pro 0813
Code migration (Vals AI)41.5%34.8%DeepSeek V4 Pro 0813
ProgramBench (Vals AI)0.0%0.0%Tie
IOI (Vals AI)51.6%67.8%Gemini 3.7 Flash
Terminal-Bench 2.1 (Vals AI)54.7%77.5%Gemini 3.7 Flash
Vibe Code Bench 1-100 (Vals AI)17.5%Not measuredOnly DeepSeek V4 Pro 0813 measured
SkillsBench (Vals AI)53.8%65.9%Gemini 3.7 Flash
CyberBench (Vals AI)72.9%43.8%DeepSeek V4 Pro 0813
SciCode51.0%59.8%Gemini 3.7 Flash
FrontierCode28.5%43.6%Gemini 3.7 Flash
FrontierSWENot measured20.3%Only Gemini 3.7 Flash measured
WeirdML66.2%Not measuredOnly DeepSeek V4 Pro 0813 measured
ALE-Bench1403.2904.3DeepSeek V4 Pro 0813
LMArena Text: coding14691499Gemini 3.7 Flash
LMArena Text: software and IT14621498Gemini 3.7 Flash
LMArena WebDev: React15881587DeepSeek V4 Pro 0813
LMArena WebDev: HTML15661593Gemini 3.7 Flash
LMArena WebDev: from an imageNot measured1599Only Gemini 3.7 Flash measured
Neither model has a result on the other 14 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
AA Intelligence Index36.039.6Gemini 3.7 Flash
Epoch Capabilities Index155.6158.0Gemini 3.7 Flash
Agentic workLike-for-likePublic boards, shown as published, kept out of the index25
Agentic work benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
AA Agentic Index41.3Not measuredOnly DeepSeek V4 Pro 0813 measured
Design Arena: agents building web appsNot measured1224Only Gemini 3.7 Flash measured
Design Arena: agents building full-stack appsNot measured1208Only Gemini 3.7 Flash measured
Design Arena: agents building mobile appsNot measured1230Only Gemini 3.7 Flash measured
Design Arena: agents building Android appsNot measured1247Only Gemini 3.7 Flash measured
Design Arena: agents building Godot gamesNot measured1264Only Gemini 3.7 Flash measured
Design Arena: agents building gamesNot measured1235Only Gemini 3.7 Flash measured
Design Arena: spreadsheetsNot measured1028Only Gemini 3.7 Flash measured
LiveBench researchNot measured76.5%Only Gemini 3.7 Flash measured
LiveBench browserNot measured56.0%Only Gemini 3.7 Flash measured
LiveBench long agenticNot measured28.3%Only Gemini 3.7 Flash measured
SREBench (Vals AI)1.9%4.6%Gemini 3.7 Flash
Finance Agent (Vals AI)50.4%59.0%Gemini 3.7 Flash
Tax Agent Bench (Vals AI)26.8%22.8%DeepSeek V4 Pro 0813
Legal Research Bench (Vals AI)40.9%34.6%DeepSeek V4 Pro 0813
Harvey's Legal Agent Benchmark (Vals AI)10.8%Not measuredOnly DeepSeek V4 Pro 0813 measured
Excel Modeling Benchmark (Vals AI)52.8%71.3%Gemini 3.7 Flash
Terminal-Bench-Science (Vals AI)4.3%7.1%Gemini 3.7 Flash
APEX-Agents47.3%67.8%Gemini 3.7 Flash
Remote Labor IndexNot measured5.0%Only Gemini 3.7 Flash measured
Neither model has a result on the other 19 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index110
Reasoning benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
GPQA Diamond92.4%93.9%Gemini 3.7 Flash
LiveBench reasoningNot measured87.8%Only Gemini 3.7 Flash measured
MMMU Pro (Vals AI)Not measured89.0%Only Gemini 3.7 Flash measured
GPQA Diamond (Artificial Analysis run)92.8%94.5%Gemini 3.7 Flash
Humanity's Last Exam (Artificial Analysis run)41.0%47.9%Gemini 3.7 Flash
GPQA Diamond (Epoch AI run)91.7%94.8%Gemini 3.7 Flash
ARC-AGI90.5%95.5%Gemini 3.7 Flash
ARC-AGI-261.3%84.6%Gemini 3.7 Flash
Chess puzzles46.0%66.0%Gemini 3.7 Flash
Mystery game puzzles43.0%37.0%DeepSeek V4 Pro 0813
LMArena Text: hard prompts14581495Gemini 3.7 Flash
LMArena Text: expert14591512Gemini 3.7 Flash
LMArena Text: hard prompts in English14631487Gemini 3.7 Flash
LMArena Vision: diagramsNot measured1329Only Gemini 3.7 Flash measured
LMArena Vision: homeworkNot measured1320Only Gemini 3.7 Flash measured
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index15
Math benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
OTIS Mock AIME98.6%97.2%DeepSeek V4 Pro 0813
FrontierMath, tiers 1 to 364.6%71.6%Gemini 3.7 Flash
FrontierMath, tier 426.8%36.6%Gemini 3.7 Flash
ProofBench50.0%58.0%Gemini 3.7 Flash
LMArena Text: math14411511Gemini 3.7 Flash
LMArena Text: mathematical work14431502Gemini 3.7 Flash
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index09
Knowledge benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
MMLU Pro87.0%90.1%Gemini 3.7 Flash
LegalBench (Vals AI)82.4%87.3%Gemini 3.7 Flash
CorpFin (Vals AI)65.4%Not measuredOnly DeepSeek V4 Pro 0813 measured
TaxEval (Vals AI)73.1%74.7%Gemini 3.7 Flash
MortgageTax (Vals AI)Not measured66.7%Only Gemini 3.7 Flash measured
MedCode (Vals AI)42.5%53.4%Gemini 3.7 Flash
SimpleQA Verified52.9%69.2%Gemini 3.7 Flash
LMArena Text: science14571511Gemini 3.7 Flash
LMArena Text: medicine and healthcare14461494Gemini 3.7 Flash
LMArena Text: legal and government14591496Gemini 3.7 Flash
LMArena Text: business and finance14421479Gemini 3.7 Flash
Neither model has a result on the other 6 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Web design benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
Design Arena: websitesNot measured1304Only Gemini 3.7 Flash measured
Design Arena: UI componentsNot measured1299Only Gemini 3.7 Flash measured
Design Arena: game devNot measured1318Only Gemini 3.7 Flash measured
Design Arena: data vizNot measured1315Only Gemini 3.7 Flash measured
Design Arena: 3DNot measured1308Only Gemini 3.7 Flash measured
Design Arena: all code categoriesNot measured1308Only Gemini 3.7 Flash measured
Design Arena: SVGNot measured1283Only Gemini 3.7 Flash measured
Design Arena: ASCII artNot measured1276Only Gemini 3.7 Flash measured
LiveBench designNot measured73.2%Only Gemini 3.7 Flash measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index018
General benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
Vals Index53.0%Not measuredOnly DeepSeek V4 Pro 0813 measured
LiveBench data analysisNot measured68.0%Only Gemini 3.7 Flash measured
SAGE (Vals AI)Not measured49.2%Only Gemini 3.7 Flash measured
MedScribe (Vals AI)80.2%83.9%Gemini 3.7 Flash
Long context reasoning (AA-LCR)80.3%83.0%Gemini 3.7 Flash
DTBench93.9%96.8%Gemini 3.7 Flash
LMCA45.5%50.4%Gemini 3.7 Flash
Furniture assemblyNot measured26.7%Only Gemini 3.7 Flash measured
LMArena Text14461489Gemini 3.7 Flash
LMArena Text: instruction following14481484Gemini 3.7 Flash
LMArena Text: creative writing14391488Gemini 3.7 Flash
LMArena VisionNot measured1314Only Gemini 3.7 Flash measured
LMArena Text: English14571483Gemini 3.7 Flash
LMArena Text: non-English14311487Gemini 3.7 Flash
LMArena Text: Chinese14811552Gemini 3.7 Flash
LMArena Text: GermanNot measured1506Only Gemini 3.7 Flash measured
LMArena Text: JapaneseNot measured1511Only Gemini 3.7 Flash measured
LMArena Text: Korean14491481Gemini 3.7 Flash
LMArena Text: French14461500Gemini 3.7 Flash
LMArena Text: Spanish14351495Gemini 3.7 Flash
LMArena Text: Russian14481515Gemini 3.7 Flash
LMArena Text: PolishNot measured1503Only Gemini 3.7 Flash measured
LMArena Text: multi-turn14541485Gemini 3.7 Flash
LMArena Text: longer queries14581492Gemini 3.7 Flash
LMArena Text: writing and language14431491Gemini 3.7 Flash
LMArena Text: entertainment and media14241477Gemini 3.7 Flash
LMArena Vision: EnglishNot measured1309Only Gemini 3.7 Flash measured
LMArena Vision: ChineseNot measured1394Only Gemini 3.7 Flash measured
LMArena Vision: reading textNot measured1322Only Gemini 3.7 Flash measured
LMArena Vision: humorNot measured1299Only Gemini 3.7 Flash measured
LMArena Vision: creative writingNot measured1312Only Gemini 3.7 Flash measured
Neither model has a result on the other 15 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
BenchmarkDeepSeek V4 Pro 0813Gemini 3.7 FlashLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

DeepSeek V4 Pro 0813$0.0016Fits in one request
Gemini 3.7 Flash$0.0026Fits in one request

DeepSeek V4 Pro 0813 costs less

Repository review

50K input + 3K output tokens

DeepSeek V4 Pro 0813$0.039Fits in one request
Gemini 3.7 Flash$0.049Fits in one request

DeepSeek V4 Pro 0813 costs less

Agent loop

200K input + 10K output tokens

DeepSeek V4 Pro 0813$0.152Fits in one request
Gemini 3.7 Flash$0.188Fits in one request

DeepSeek V4 Pro 0813 costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of DeepSeek V4 Pro 0813 and Gemini 3.7 Flash
SpecificationDeepSeek V4 Pro 0813Google Gemini 3.7 Flash
MakerDeepSeekGoogle
WeightsOpen weightsNo public weights listed
ReleasedAugust 12, 2026August 13, 2026
Context window1.05M1.05M
Input $/Mtok$0.66$0.75
Output $/Mtok$1.98$3.75
Coding index54.669.9
Independent measurements behind it12
Cost per solved task$0.10$1.78
SourceOpenRouter listingOpenRouter listing

Questions

Is DeepSeek V4 Pro 0813 better than Gemini 3.7 Flash for coding?

Pick Gemini 3.7 Flash for coding: it scores 69.9 on the Vibe Coding Index against 54.6 for DeepSeek V4 Pro 0813. Pick DeepSeek V4 Pro 0813 when output tokens cost 1.9x less, or when a solved task costs $0.10 against $1.78.

Which is cheaper, DeepSeek V4 Pro 0813 or Gemini 3.7 Flash?

DeepSeek V4 Pro 0813 costs $0.66 per million input tokens and $1.98 per million output tokens; Gemini 3.7 Flash costs $0.75 and $3.75. Per solved SWE-bench Verified task, DeepSeek V4 Pro 0813 costs $0.10 and Gemini 3.7 Flash $1.78.

Which has the bigger context window, DeepSeek V4 Pro 0813 or Gemini 3.7 Flash?

Neither: both take 1.05M tokens.

How many benchmarks is the DeepSeek V4 Pro 0813 vs Gemini 3.7 Flash comparison based on?

72 benchmarks have a result for both models as of October 21, 2026. DeepSeek V4 Pro 0813's coding index rests on 1 independent benchmark and Gemini 3.7 Flash's on 2. Where only one model has a result, the page says so instead of naming a winner.

Compare next