LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

DeepSeek V4 Pro 0423 vs Gemini 3.5 Flash

Updated October 21, 2026. 78 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 0423 Replace Gemini 3.5 Flash
DeepSeek V4 Pro 0423DeepSeek
Coding
35.3
Overall
30.4
Output $/M
$1.9105
7
Gemini 3.5 FlashGoogle
Coding
42.8
Overall
33.6
Output $/M
$9.00
70
vs boards won of 78 shared, 1 tied

Decision reading

Pick Gemini 3.5 Flash for coding: it scores 42.8 on the Vibe Coding Index against 35.3 for DeepSeek V4 Pro 0423.

Pick DeepSeek V4 Pro 0423 when output tokens cost 4.7x less, or when a solved task costs $0.57 against $1.21.

Model ADeepSeek V4 Pro 0423DeepSeek35.3/1001 independent benchmark$0.57 per solved taskModel BGemini 3.5 FlashGoogle42.8/1001 independent benchmark$1.21 per solved task
Shared results
78
DeepSeek V4 Pro 0423 only
5
Gemini 3.5 Flash only
32
Like-for-like categories
8 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 0423 35.3 from 1 benchmark, Gemini 3.5 Flash 42.8 from 1 benchmark.
Widest category
Agentic work, 13.9 points apart on average over 8 shared benchmarks.
Widest single gap
Design Arena: agents building full-stack apps: DeepSeek V4 Pro 0423 948 against Gemini 3.5 Flash 1197, 31.1 points apart on a 0 to 100 scale.
Output price
DeepSeek V4 Pro 0423 $1.9105 against Gemini 3.5 Flash $9.00 per million output tokens, 4.7x 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 0423BarsGemini 3.5 FlashTies
Coding3111 tie
Intelligence02
Agentic work17
Reasoning07
Math15
Knowledge010
Web design17
General121
All shared boards7701 tie
CodingLike-for-likeFour of these boards feed the Vibe Coding Index311
Coding benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 FlashLeader
LMArena WebDev14471499Gemini 3.5 Flash
AA Coding Index59.4Not measuredOnly DeepSeek V4 Pro 0423 measured
LiveCodeBench (Vals AI)87.5%87.6%Gemini 3.5 Flash
SWE-bench Verified (Vals AI)77.4%78.8%Gemini 3.5 Flash
Terminal-Bench 4 (Vals AI)11.1%6.1%DeepSeek V4 Pro 0423
Vibe Code Bench (Vals AI)49.9%48.7%DeepSeek V4 Pro 0423
Code migration (Vals AI)26.2%26.8%Gemini 3.5 Flash
ProgramBench (Vals AI)0.0%0.0%Tie
Terminal-Bench 2.1 (Vals AI)50.2%74.2%Gemini 3.5 Flash
SkillsBench (Vals AI)51.3%52.7%Gemini 3.5 Flash
SciCodeNot measured53.9%Only Gemini 3.5 Flash measured
Terminal-Bench Hard (Artificial Analysis run)Not measured46.2%Only Gemini 3.5 Flash measured
FrontierCode17.6%Not measuredOnly DeepSeek V4 Pro 0423 measured
WeirdML48.9%62.6%Gemini 3.5 Flash
ALE-Bench1006.1911.0DeepSeek V4 Pro 0423
LMArena Text: coding14701493Gemini 3.5 Flash
LMArena Text: software and IT14691494Gemini 3.5 Flash
LMArena WebDev: React14411495Gemini 3.5 Flash
LMArena WebDev: HTML14351522Gemini 3.5 Flash
LMArena WebDev: from an imageNot measured1504Only Gemini 3.5 Flash measured
Neither model has a result on the other 18 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 FlashLeader
AA Intelligence Index30.433.6Gemini 3.5 Flash
Epoch Capabilities Index149.1154.9Gemini 3.5 Flash
Agentic workLike-for-likePublic boards, shown as published, kept out of the index17
Agentic work benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 FlashLeader
AA Agentic Index26.3Not measuredOnly DeepSeek V4 Pro 0423 measured
Design Arena: agents building web apps10001202Gemini 3.5 Flash
Design Arena: agents building full-stack apps9481197Gemini 3.5 Flash
Design Arena: agents building mobile appsNot measured1175Only Gemini 3.5 Flash measured
Design Arena: agents building Android appsNot measured1190Only Gemini 3.5 Flash measured
Design Arena: agents building Godot games10591098Gemini 3.5 Flash
Design Arena: agents building gamesNot measured1180Only Gemini 3.5 Flash measured
Design Arena: HTML slidesNot measured1142Only Gemini 3.5 Flash measured
Design Arena: PowerPoint slidesNot measured1178Only Gemini 3.5 Flash measured
Finance Agent (Vals AI)44.1%57.9%Gemini 3.5 Flash
Tax Agent Bench (Vals AI)25.2%21.6%DeepSeek V4 Pro 0423
Legal Research Bench (Vals AI)23.1%30.8%Gemini 3.5 Flash
Excel Modeling Benchmark (Vals AI)51.6%63.5%Gemini 3.5 Flash
Terminal-Bench-Science (Vals AI)Not measured5.7%Only Gemini 3.5 Flash measured
τ²-Bench TelecomNot measured95.6%Only Gemini 3.5 Flash measured
APEX-AgentsNot measured27.5%Only Gemini 3.5 Flash measured
Vending-Bench 2$3,285$5,396Gemini 3.5 Flash
Neither model has a result on the other 22 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index07
Reasoning benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 FlashLeader
GPQA Diamond89.4%92.7%Gemini 3.5 Flash
MMMU Pro (Vals AI)Not measured88.3%Only Gemini 3.5 Flash measured
GPQA Diamond (Artificial Analysis run)Not measured92.2%Only Gemini 3.5 Flash measured
Humanity's Last Exam (Artificial Analysis run)Not measured42.7%Only Gemini 3.5 Flash measured
GPQA Diamond (Epoch AI run)90.9%92.8%Gemini 3.5 Flash
ARC-AGINot measured92.5%Only Gemini 3.5 Flash measured
ARC-AGI-2Not measured72.1%Only Gemini 3.5 Flash measured
SimpleBenchNot measured76.7%Only Gemini 3.5 Flash measured
Chess puzzles19.0%45.0%Gemini 3.5 Flash
Mystery game puzzles17.0%32.0%Gemini 3.5 Flash
EBR-benchNot measured4.8%Only Gemini 3.5 Flash measured
LMArena Text: hard prompts14611488Gemini 3.5 Flash
LMArena Text: expert14631495Gemini 3.5 Flash
LMArena Text: hard prompts in English14661483Gemini 3.5 Flash
LMArena Vision: diagramsNot measured1323Only Gemini 3.5 Flash measured
LMArena Vision: homeworkNot measured1330Only Gemini 3.5 Flash measured
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index15
Math benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 FlashLeader
OTIS Mock AIME96.7%95.6%DeepSeek V4 Pro 0423
FrontierMath, tiers 1 to 345.3%62.8%Gemini 3.5 Flash
FrontierMath, tier 42.4%26.8%Gemini 3.5 Flash
ProofBench16.0%31.0%Gemini 3.5 Flash
LMArena Text: math14431504Gemini 3.5 Flash
LMArena Text: mathematical work14551495Gemini 3.5 Flash
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index010
Knowledge benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 FlashLeader
MMLU Pro87.3%89.5%Gemini 3.5 Flash
LegalBench (Vals AI)80.3%83.6%Gemini 3.5 Flash
CaseLaw (Vals AI)59.4%Not measuredOnly DeepSeek V4 Pro 0423 measured
CorpFin (Vals AI)61.4%64.7%Gemini 3.5 Flash
TaxEval (Vals AI)72.1%74.4%Gemini 3.5 Flash
MortgageTax (Vals AI)Not measured68.1%Only Gemini 3.5 Flash measured
MedCode (Vals AI)40.5%55.8%Gemini 3.5 Flash
SimpleQA Verified47.0%66.2%Gemini 3.5 Flash
LMArena Text: science14701495Gemini 3.5 Flash
LMArena Text: medicine and healthcare14661483Gemini 3.5 Flash
LMArena Text: legal and government14641488Gemini 3.5 Flash
LMArena Text: business and finance14431467Gemini 3.5 Flash
Neither model has a result on the other 5 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index17
Web design benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 FlashLeader
Design Arena: websites12351268Gemini 3.5 Flash
Design Arena: UI components12211272Gemini 3.5 Flash
Design Arena: game dev12291263Gemini 3.5 Flash
Design Arena: data viz12031236Gemini 3.5 Flash
Design Arena: 3D12541244DeepSeek V4 Pro 0423
Design Arena: all code categories12401266Gemini 3.5 Flash
Design Arena: SVG11391254Gemini 3.5 Flash
Design Arena: ASCII art11481253Gemini 3.5 Flash
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index121
General benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 FlashLeader
Public Benefits Bench (Vals AI)62.9%59.5%DeepSeek V4 Pro 0423
SAGE (Vals AI)Not measured49.9%Only Gemini 3.5 Flash measured
MedScribe (Vals AI)75.1%76.6%Gemini 3.5 Flash
Vals Multimodal IndexNot measured62.9%Only Gemini 3.5 Flash measured
IFBenchNot measured76.3%Only Gemini 3.5 Flash measured
Long context reasoning (AA-LCR)Not measured74.3%Only Gemini 3.5 Flash measured
DTBench90.7%94.7%Gemini 3.5 Flash
LMCA41.2%47.1%Gemini 3.5 Flash
Surface Evolver Bench40.0%58.1%Gemini 3.5 Flash
CL-bench Life13.5%Not measuredOnly DeepSeek V4 Pro 0423 measured
LMArena Text14511482Gemini 3.5 Flash
LMArena Text: instruction following14451467Gemini 3.5 Flash
LMArena Text: creative writing14461470Gemini 3.5 Flash
LMArena VisionNot measured1308Only Gemini 3.5 Flash measured
LMArena Text: English14591481Gemini 3.5 Flash
LMArena Text: non-English14391475Gemini 3.5 Flash
LMArena Text: Chinese14811527Gemini 3.5 Flash
LMArena Text: German14581494Gemini 3.5 Flash
LMArena Text: Japanese14461486Gemini 3.5 Flash
LMArena Text: Korean14091448Gemini 3.5 Flash
LMArena Text: French14731490Gemini 3.5 Flash
LMArena Text: Spanish14571480Gemini 3.5 Flash
LMArena Text: Russian14531493Gemini 3.5 Flash
LMArena Text: Polish14611516Gemini 3.5 Flash
LMArena Text: multi-turn14681481Gemini 3.5 Flash
LMArena Text: longer queries14581483Gemini 3.5 Flash
LMArena Text: writing and language14471478Gemini 3.5 Flash
LMArena Text: entertainment and media14311459Gemini 3.5 Flash
LMArena Vision: EnglishNot measured1305Only Gemini 3.5 Flash measured
LMArena Vision: ChineseNot measured1377Only Gemini 3.5 Flash measured
LMArena Vision: reading textNot measured1317Only Gemini 3.5 Flash measured
LMArena Vision: humorNot measured1321Only Gemini 3.5 Flash measured
LMArena Vision: creative writingNot measured1301Only Gemini 3.5 Flash measured
LMArena DocumentNot measured1463Only Gemini 3.5 Flash measured
Neither model has a result on the other 12 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for DeepSeek V4 Pro 0423 and Gemini 3.5 Flash
BenchmarkDeepSeek V4 Pro 0423Gemini 3.5 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 0423$0.0019Fits in one request
Gemini 3.5 Flash$0.0060Fits in one request

DeepSeek V4 Pro 0423 costs less

Repository review

50K input + 3K output tokens

DeepSeek V4 Pro 0423$0.053Fits in one request
Gemini 3.5 Flash$0.102Fits in one request

DeepSeek V4 Pro 0423 costs less

Agent loop

200K input + 10K output tokens

DeepSeek V4 Pro 0423$0.210Fits in one request
Gemini 3.5 Flash$0.390Fits in one request

DeepSeek V4 Pro 0423 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 0423 and Gemini 3.5 Flash
SpecificationDeepSeek V4 Pro 0423Google Gemini 3.5 Flash
MakerDeepSeekGoogle
WeightsOpen weightsNo public weights listed
ReleasedApril 24, 2026May 19, 2026
Context window1.05M1.05M
Input $/Mtok$0.9553$1.50
Output $/Mtok$1.9105$9.00
Coding index35.342.8
Independent measurements behind it11
Cost per solved task$0.57$1.21
SourceOpenRouter listingOpenRouter listing

Questions

Is DeepSeek V4 Pro 0423 better than Gemini 3.5 Flash for coding?

Pick Gemini 3.5 Flash for coding: it scores 42.8 on the Vibe Coding Index against 35.3 for DeepSeek V4 Pro 0423. Pick DeepSeek V4 Pro 0423 when output tokens cost 4.7x less, or when a solved task costs $0.57 against $1.21.

Which is cheaper, DeepSeek V4 Pro 0423 or Gemini 3.5 Flash?

DeepSeek V4 Pro 0423 costs $0.9553 per million input tokens and $1.9105 per million output tokens; Gemini 3.5 Flash costs $1.50 and $9.00. Per solved SWE-bench Verified task, DeepSeek V4 Pro 0423 costs $0.57 and Gemini 3.5 Flash $1.21.

Which has the bigger context window, DeepSeek V4 Pro 0423 or Gemini 3.5 Flash?

Neither: both take 1.05M tokens.

How many benchmarks is the DeepSeek V4 Pro 0423 vs Gemini 3.5 Flash comparison based on?

78 benchmarks have a result for both models as of October 21, 2026. DeepSeek V4 Pro 0423's coding index rests on 1 independent benchmark and Gemini 3.5 Flash's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next