DeepSeek V4 Flash 0731 vs Gemini 3.5 Flash Lite

DeepSeek V4 Flash 0731 leads Gemini 3.5 Flash Lite on coding, 51.8 to 49.0 on the Vibe Coding Index, and DeepSeek V4 Flash 0731 costs 8.9x less per output token ($0.28 vs $2.50 per million). Updated October 21, 2026; 39 benchmarks have a result for both.

Replace DeepSeek V4 Flash 0731 Replace Gemini 3.5 Flash Lite

DeepSeek V4 Flash 0731DeepSeek

Coding: 51.8
Overall: 55.2
Output $/M: $0.28

Gemini 3.5 Flash LiteGoogle

Coding: 49.0
Overall: 49.8
Output $/M: $2.50

vs boards won of 39 shared, 1 tied

DeepSeek V4 Flash 0731 or Gemini 3.5 Flash Lite: which to pick

Pick DeepSeek V4 Flash 0731 for coding: it scores 51.8 on the Vibe Coding Index against 49.0 for Gemini 3.5 Flash Lite.

Pick Gemini 3.5 Flash Lite for work measured by LegalBench (Vals AI) or TaxEval (Vals AI), where it leads.

Model ADeepSeek V4 Flash 0731DeepSeek51.8/1004 independent benchmarks$0.01 per solved task Model BGemini 3.5 Flash LiteGoogle49.0/1003 independent benchmarks$0.41 per solved task

391646

Shared results: 39
DeepSeek V4 Flash 0731 only: 16
Gemini 3.5 Flash Lite only: 46
Like-for-like categories: 7 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

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.

Coding
Vibe Coding Index, six public coding benchmarks
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 scores 51.8 against 49.0.
4 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 leads, 34.3 against 22.2.
Same board, both measured
Agentic work
AA Agentic Index
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 leads, 41.0 against 14.3.
Same board, both measured
Reasoning
GPQA Diamond
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 leads, 89.9% against 83.8%.
Same board, both measured
Knowledge
MMLU Pro
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 leads, 86.2% against 85.8%.
Same board, both measured
Web design
Design Arena: websites
No call
Only DeepSeek V4 Flash 0731 has a Design Arena: websites result, 1238.
Directional only
Long documents
Listed context window
No call
Both take 1.05M tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 $0.0003, Gemini 3.5 Flash Lite $0.0015 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 $0.01, Gemini 3.5 Flash Lite $0.41.
Published attempt cost and pass rate

At a glance

Coding index: DeepSeek V4 Flash 0731 51.8 from 4 benchmarks, Gemini 3.5 Flash Lite 49.0 from 3 benchmarks.
Widest category: Math, 30.6 points apart on average over 4 shared benchmarks.
Widest single gap: ARC-AGI-2: DeepSeek V4 Flash 0731 61.4% against Gemini 3.5 Flash Lite 10.3%, 51.1 points apart on a 0 to 100 scale.
Output price: DeepSeek V4 Flash 0731 $0.28 against Gemini 3.5 Flash Lite $2.50 per million output tokens, 8.9x 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
Category	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Ties
Coding	10	0	1 tie
Intelligence	2	0
Agentic work	5	0
Reasoning	8	0
Math	4	0
Knowledge	2	3
General	4	0
All shared boards	35	3	1 tie

CodingLike-for-likeSix of these boards feed the Vibe Coding Index100

Coding benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
LMArena WebDev	Not measured	1435	Only Gemini 3.5 Flash Lite measured
AA Coding Index	69.1	49.3	DeepSeek V4 Flash 0731
LiveCodeBench (Vals AI)	87.3%	79.0%	DeepSeek V4 Flash 0731
SWE-bench Verified (Vals AI)	88.8%	75.0%	DeepSeek V4 Flash 0731
Terminal-Bench 4 (Vals AI)	18.7%	Not measured	Only DeepSeek V4 Flash 0731 measured
Vibe Code Bench (Vals AI)	74.7%	37.2%	DeepSeek V4 Flash 0731
Code migration (Vals AI)	38.6%	6.2%	DeepSeek V4 Flash 0731
ProgramBench (Vals AI)	0.0%	0.0%	Tie
IOI (Vals AI)	32.7%	Not measured	Only DeepSeek V4 Flash 0731 measured
Terminal-Bench 2.1 (Vals AI)	67.0%	50.2%	DeepSeek V4 Flash 0731
SkillsBench (Vals AI)	50.7%	Not measured	Only DeepSeek V4 Flash 0731 measured
CyberBench (Vals AI)	72.1%	59.1%	DeepSeek V4 Flash 0731
SciCode	50.3%	41.3%	DeepSeek V4 Flash 0731
FrontierCode	18.8%	Not measured	Only DeepSeek V4 Flash 0731 measured
WeirdML	63.0%	39.0%	DeepSeek V4 Flash 0731
ALE-Bench	1306.1	765.3	DeepSeek V4 Flash 0731
LMArena Text: coding	Not measured	1453	Only Gemini 3.5 Flash Lite measured
LMArena Text: software and IT	Not measured	1451	Only Gemini 3.5 Flash Lite measured
LMArena WebDev: React	Not measured	1408	Only Gemini 3.5 Flash Lite measured
LMArena WebDev: from an image	Not measured	1448	Only Gemini 3.5 Flash Lite measured
SciCode	Not measured	6.2%	Only Gemini 3.5 Flash Lite measured
Neither model has a result on the other 18 boards here

IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20

Intelligence benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
AA Intelligence Index	34.3	22.2	DeepSeek V4 Flash 0731
Epoch Capabilities Index	154.8	144.9	DeepSeek V4 Flash 0731

Agentic workLike-for-likePublic boards, shown as published, kept out of the index50

Agentic work benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
AA Agentic Index	41.0	14.3	DeepSeek V4 Flash 0731
Finance Agent (Vals AI)	49.5%	47.4%	DeepSeek V4 Flash 0731
Tax Agent Bench (Vals AI)	28.1%	7.2%	DeepSeek V4 Flash 0731
Legal Research Bench (Vals AI)	30.3%	13.9%	DeepSeek V4 Flash 0731
Excel Modeling Benchmark (Vals AI)	57.0%	43.2%	DeepSeek V4 Flash 0731
BioMysteryBench (Vals AI)	64.4%	Not measured	Only DeepSeek V4 Flash 0731 measured
Terminal-Bench-Science (Vals AI)	4.3%	Not measured	Only DeepSeek V4 Flash 0731 measured
APEX-Agents	Not measured	29.3%	Only Gemini 3.5 Flash Lite measured
Neither model has a result on the other 33 boards here

ReasoningLike-for-likePublic boards, shown as published, kept out of the index80

Reasoning benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
GPQA Diamond	89.9%	83.8%	DeepSeek V4 Flash 0731
MMMU Pro (Vals AI)	Not measured	83.6%	Only Gemini 3.5 Flash Lite measured
GPQA Diamond (Artificial Analysis run)	90.8%	83.8%	DeepSeek V4 Flash 0731
Humanity's Last Exam (Artificial Analysis run)	38.6%	18.8%	DeepSeek V4 Flash 0731
GPQA Diamond (Epoch AI run)	91.0%	83.3%	DeepSeek V4 Flash 0731
ARC-AGI	89.0%	53.5%	DeepSeek V4 Flash 0731
ARC-AGI-2	61.4%	10.3%	DeepSeek V4 Flash 0731
SimpleBench	61.1%	Not measured	Only DeepSeek V4 Flash 0731 measured
Chess puzzles	32.0%	21.0%	DeepSeek V4 Flash 0731
Mystery game puzzles	34.0%	19.0%	DeepSeek V4 Flash 0731
LMArena Text: hard prompts	Not measured	1441	Only Gemini 3.5 Flash Lite measured
LMArena Text: expert	Not measured	1434	Only Gemini 3.5 Flash Lite measured
LMArena Text: hard prompts in English	Not measured	1440	Only Gemini 3.5 Flash Lite measured
LMArena Vision: diagrams	Not measured	1284	Only Gemini 3.5 Flash Lite measured
LMArena Vision: homework	Not measured	1291	Only Gemini 3.5 Flash Lite measured
Neither model has a result on the other 7 boards here

MathLike-for-likePublic boards, shown as published, kept out of the index40

Math benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
OTIS Mock AIME	94.4%	71.1%	DeepSeek V4 Flash 0731
FrontierMath, tiers 1 to 3	57.5%	26.0%	DeepSeek V4 Flash 0731
FrontierMath, tier 4	24.4%	0.0%	DeepSeek V4 Flash 0731
ProofBench	56.0%	13.0%	DeepSeek V4 Flash 0731
LMArena Text: math	Not measured	1434	Only Gemini 3.5 Flash Lite measured
LMArena Text: mathematical work	Not measured	1433	Only Gemini 3.5 Flash Lite measured
Neither model has a result on the other 7 boards here

KnowledgeLike-for-likePublic boards, shown as published, kept out of the index23

Knowledge benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
MMLU Pro	86.2%	85.8%	DeepSeek V4 Flash 0731
LegalBench (Vals AI)	77.7%	84.1%	Gemini 3.5 Flash Lite
CorpFin (Vals AI)	61.9%	60.7%	DeepSeek V4 Flash 0731
TaxEval (Vals AI)	70.7%	72.6%	Gemini 3.5 Flash Lite
MortgageTax (Vals AI)	Not measured	68.7%	Only Gemini 3.5 Flash Lite measured
MedCode (Vals AI)	41.4%	43.5%	Gemini 3.5 Flash Lite
SimpleQA Verified	33.6%	Not measured	Only DeepSeek V4 Flash 0731 measured
LMArena Text: science	Not measured	1451	Only Gemini 3.5 Flash Lite measured
LMArena Text: medicine and healthcare	Not measured	1438	Only Gemini 3.5 Flash Lite measured
LMArena Text: legal and government	Not measured	1431	Only Gemini 3.5 Flash Lite measured
LMArena Text: business and finance	Not measured	1423	Only Gemini 3.5 Flash Lite measured
FACTS Benchmark Suite	Not measured	53.5%	Only Gemini 3.5 Flash Lite measured
Neither model has a result on the other 6 boards here

Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared

Web design benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
Design Arena: websites	1238	Not measured	Only DeepSeek V4 Flash 0731 measured
Design Arena: UI components	1235	Not measured	Only DeepSeek V4 Flash 0731 measured
Design Arena: game dev	1208	Not measured	Only DeepSeek V4 Flash 0731 measured
Design Arena: data viz	1180	Not measured	Only DeepSeek V4 Flash 0731 measured
Design Arena: 3D	1209	Not measured	Only DeepSeek V4 Flash 0731 measured
Design Arena: all code categories	1230	Not measured	Only DeepSeek V4 Flash 0731 measured
Design Arena: SVG	1185	Not measured	Only DeepSeek V4 Flash 0731 measured
Design Arena: ASCII art	1090	Not measured	Only DeepSeek V4 Flash 0731 measured
Neither model has a result on the other 1 board here

GeneralLike-for-likePublic boards, shown as published, kept out of the index40

General benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
Public Benefits Bench (Vals AI)	Not measured	51.8%	Only Gemini 3.5 Flash Lite measured
SAGE (Vals AI)	Not measured	47.3%	Only Gemini 3.5 Flash Lite measured
MedScribe (Vals AI)	80.4%	70.9%	DeepSeek V4 Flash 0731
Long context reasoning (AA-LCR)	79.7%	76.0%	DeepSeek V4 Flash 0731
DTBench	90.9%	83.5%	DeepSeek V4 Flash 0731
LMCA	41.7%	37.6%	DeepSeek V4 Flash 0731
LMArena Text	Not measured	1435	Only Gemini 3.5 Flash Lite measured
LMArena Text: instruction following	Not measured	1419	Only Gemini 3.5 Flash Lite measured
LMArena Text: creative writing	Not measured	1421	Only Gemini 3.5 Flash Lite measured
LMArena Vision	Not measured	1269	Only Gemini 3.5 Flash Lite measured
LMArena Text: English	Not measured	1438	Only Gemini 3.5 Flash Lite measured
LMArena Text: non-English	Not measured	1427	Only Gemini 3.5 Flash Lite measured
LMArena Text: Chinese	Not measured	1468	Only Gemini 3.5 Flash Lite measured
LMArena Text: German	Not measured	1454	Only Gemini 3.5 Flash Lite measured
LMArena Text: Japanese	Not measured	1426	Only Gemini 3.5 Flash Lite measured
LMArena Text: Korean	Not measured	1407	Only Gemini 3.5 Flash Lite measured
LMArena Text: French	Not measured	1451	Only Gemini 3.5 Flash Lite measured
LMArena Text: Spanish	Not measured	1441	Only Gemini 3.5 Flash Lite measured
LMArena Text: Russian	Not measured	1443	Only Gemini 3.5 Flash Lite measured
LMArena Text: Polish	Not measured	1437	Only Gemini 3.5 Flash Lite measured
LMArena Text: multi-turn	Not measured	1445	Only Gemini 3.5 Flash Lite measured
LMArena Text: longer queries	Not measured	1435	Only Gemini 3.5 Flash Lite measured
LMArena Text: writing and language	Not measured	1428	Only Gemini 3.5 Flash Lite measured
LMArena Text: entertainment and media	Not measured	1414	Only Gemini 3.5 Flash Lite measured
LMArena Vision: English	Not measured	1269	Only Gemini 3.5 Flash Lite measured
LMArena Vision: Chinese	Not measured	1311	Only Gemini 3.5 Flash Lite measured
LMArena Vision: reading text	Not measured	1280	Only Gemini 3.5 Flash Lite measured
LMArena Vision: humor	Not measured	1287	Only Gemini 3.5 Flash Lite measured
LMArena Vision: creative writing	Not measured	1260	Only Gemini 3.5 Flash Lite measured
Neither model has a result on the other 17 boards here

Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared

Image generation benchmarks for DeepSeek V4 Flash 0731 and Gemini 3.5 Flash Lite
Benchmark	DeepSeek V4 Flash 0731	Gemini 3.5 Flash Lite	Leader
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 Flash 0731$0.0003Fits in one request

Gemini 3.5 Flash Lite$0.0015Fits in one request

DeepSeek V4 Flash 0731 costs less

Repository review

50K input + 3K output tokens

DeepSeek V4 Flash 0731$0.0078Fits in one request

Gemini 3.5 Flash Lite$0.022Fits in one request

DeepSeek V4 Flash 0731 costs less

Agent loop

200K input + 10K output tokens

DeepSeek V4 Flash 0731$0.031Fits in one request

Gemini 3.5 Flash Lite$0.085Fits in one request

DeepSeek V4 Flash 0731 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 Flash 0731 and Gemini 3.5 Flash Lite
Specification	DeepSeek V4 Flash 0731	Google Gemini 3.5 Flash Lite
Maker	DeepSeek	Google
Weights	Open weights	No public weights listed
Released	July 31, 2026	July 21, 2026
Context window	1.05M	1.05M
Input $/Mtok	$0.14	$0.30
Output $/Mtok	$0.28	$2.50
Coding index	51.8	49.0
Independent measurements behind it	4	3
Cost per solved task	$0.01	$0.41
Source	OpenRouter listing	OpenRouter listing

Questions

Is DeepSeek V4 Flash 0731 better than Gemini 3.5 Flash Lite for coding?

Pick DeepSeek V4 Flash 0731 for coding: it scores 51.8 on the Vibe Coding Index against 49.0 for Gemini 3.5 Flash Lite. Pick Gemini 3.5 Flash Lite for work measured by LegalBench (Vals AI) or TaxEval (Vals AI), where it leads.

Which is cheaper, DeepSeek V4 Flash 0731 or Gemini 3.5 Flash Lite?

DeepSeek V4 Flash 0731 costs $0.14 per million input tokens and $0.28 per million output tokens; Gemini 3.5 Flash Lite costs $0.30 and $2.50. Per solved SWE-bench Verified task, DeepSeek V4 Flash 0731 costs $0.01 and Gemini 3.5 Flash Lite $0.41.

Which has the bigger context window, DeepSeek V4 Flash 0731 or Gemini 3.5 Flash Lite?

Neither: both take 1.05M tokens.

How many benchmarks is the DeepSeek V4 Flash 0731 vs Gemini 3.5 Flash Lite comparison based on?

39 benchmarks have a result for both models as of October 21, 2026. DeepSeek V4 Flash 0731's coding index rests on 4 independent benchmarks and Gemini 3.5 Flash Lite's on 3. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/deepseek-v4-flash-0731-vs-gemini-3-5-flash-lite" title="DeepSeek V4 Flash 0731 vs Gemini 3.5 Flash Lite, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/deepseek-v4-flash-0731-vs-gemini-3-5-flash-lite">LLMBenchmark.ai, DeepSeek V4 Flash 0731 vs Gemini 3.5 Flash Lite</a></p>

DeepSeek V4 Flash 0731 or Gemini 3.5 Flash Lite: which to pick

Which one for your work

Coding

Intelligence

Agentic work

Reasoning

Knowledge

Web design

Long documents

Everyday cost

Cost per solved task

At a glance

Benchmark by benchmark

What each workload costs

Chat turn

Repository review

Agent loop

Specification differences

Questions

Compare next