OpenAIDeepSeekLLM matchup

GPT-5.6 Lunavs.DeepSeek-V4.1-Flash

Which one is better?

DeepSeek-V4.1-Flash performs better than GPT-5.6 Luna and is cheaper: GPT-5.6 Luna costs 1.3 times (input) and 2 times (output) as much.

DeepSeek-V4.1-Flash ahead

IntelligenceOverall index, 100 points = average

93.9

GPT-5.6 Luna

Rank 38

99.7

DeepSeek-V4.1-Flash

Rank 29

DeepSeek-V4.1-Flash faster

SpeedTokens per second

GPT-5.6 Luna

176

DeepSeek-V4.1-Flash

DeepSeek-V4.1-Flash cheaper

Price$ per 1M tokens

$0.45

GPT-5.6 Luna

$0.26

DeepSeek-V4.1-Flash

GPT-5.6 Luna more popular

Popularity0 to 100

GPT-5.6 Luna

Rank 161 of 875

DeepSeek-V4.1-Flash

Rank 298 of 875

Above 272,000 tokens of context, GPT-5.6 Luna costs more: $0.40 input and $1.80 output per 1M tokens.

Choose GPT-5.6 Luna for

No clear advantage

Choose DeepSeek-V4.1-Flash for

AgentsProductivityLegalCybersecurityMultimodalGeneral (not certain)Finance (not certain)Health (not certain)

Where is the difference?

Strengths at a glance

GPT-5.6 LunaDeepSeek-V4.1-Flash

The median of all ranked models is below 100 in every area, so it sits in the centre of the net.

Codinglevel

GPT-5.6 Luna96.0Rank 40

DeepSeek-V4.1-Flash98.8Rank 30

Median of ranked models83.7

95 % range: GPT-5.6 Luna 86.4 to 105.6, DeepSeek-V4.1-Flash 91.4 to 106.3.

AreaGPT-5.6 LunaDeepSeek-V4.1-Flash

The same values and verdicts as the bars below. In the top field the net starts at 100 points, the reference group average; values below sit as a hollow point in the centre. Not in the net because a model has no own value there or the data are too thin: Reasoning, Mathematics, Knowledge.

DeepSeek-V4.1-Flash is ahead in 3 of 10 areas, GPT-5.6 Luna in none.

level: 1too little data: 3DeepSeek-V4.1-Flash slightly ahead: 3DeepSeek-V4.1-Flash ahead: 3

Areas in the overall index

AgentsDeepSeek-V4.1-Flash +997.6 to 106.4

MultimodalDeepSeek-V4.1-Flash +492.2 to 95.7

Generalleaning DeepSeek-V4.1-Flashnot certain · 5 and 4 tests

Codinglevel96.0 to 98.8

Mathematicstoo little data94.9 to 97.3

Reasoningonly GPT-5.6 Luna has a score

Knowledgeonly GPT-5.6 Luna has a score

Specialist fields

LegalDeepSeek-V4.1-Flash +695.2 to 100.8

Financeleaning DeepSeek-V4.1-Flashnot certain · 6 and 5 tests

Healthleaning DeepSeek-V4.1-Flashnot certain · 2 tests each

Bar length shows the gap in index points (full width 15 points, one standard deviation of the reference group). Solid means clearly ahead, hatched not certain yet, a thin grey bar level.

Show 3 more areas that do not count toward the overall index

ProductivityDeepSeek-V4.1-Flash +792.5 to 99.7

CybersecurityDeepSeek-V4.1-Flash +598.0 to 103.1

Educationonly DeepSeek-V4.1-Flash has a score

Bar length shows the gap in index points (full width 15 points, one standard deviation of the reference group). Solid means clearly ahead, hatched not certain yet, a thin grey bar level.

GPT-5.6 Luna vs. DeepSeek-V4.1-Flash at the same effort

Both models measured at the same reasoning efforts. This shows whether a lead comes from the model or only from more compute.

maxDeepSeek-V4.1-Flash slightly ahead

GPT-5.6 Luna 93.9 vs. DeepSeek-V4.1-Flash 99.7 · Gap 3.2 · Test series 3

Overall index (max vs. max)DeepSeek-V4.1-Flash ahead

GPT-5.6 Luna 93.9 vs. DeepSeek-V4.1-Flash 99.7

Effort	GPT-5.6 Luna	DeepSeek-V4.1-Flash	Gap	Verdict	Test series
max	93.9	99.7	3.2	DeepSeek-V4.1-Flash slightly ahead	3
Overall index (max vs. max)	93.9	99.7	5.8	DeepSeek-V4.1-Flash ahead

Gap and verdict compare only the Artificial Analysis test series in which both models are measured at that effort. Tied below 3 points, ahead only beyond two standard errors. The single values rest on all test series of each model, so their difference can deviate from the gap.

Which model for which task?

Everyday tasks, translated from the area values above. Each row names the values it rests on.

TaskBetter choicebased on

Writing and debugging codelevelCoding 96.0 to 98.8

Maths and data analysisopenMathematics 94.9 to 97.3 (too little data), Reasoning without a score for DeepSeek-V4.1-Flash

Facts and researchopenKnowledge without a score for DeepSeek-V4.1-Flash

Agents and workflowsDeepSeek-V4.1-FlashAgents 97.6 to 106.4

Fast chat repliesDeepSeek-V4.1-Flash94 to 176 tokens per second

Security analysisDeepSeek-V4.1-FlashCybersecurity 98.0 to 103.1

Many requests on a small budgetDeepSeek-V4.1-Flash$0.45 to $0.26 per 1M tokens

Finance, legal and medical textsDeepSeek-V4.1-FlashFinance 98.2 to 103.7, Legal 95.2 to 100.8, Health 92.4 to 95.5

Images, screenshots and PDFsDeepSeek-V4.1-FlashMultimodal 92.2 to 95.7

See for yourself: same task, one attempt

The same prompt for both models. We show the latest attempt without fixes; it does not count toward the overall index.

Pelican on a bicycle

Prompt

Generate an SVG of a pelican riding a bicycle

Task origin: Simon Willison

GPT-5.6 Luna

Sep 7, 2026 · 50 s · $0.0038

Test details

Run date: Sep 7, 2026
API provider: openrouter
Requested API model: openai/gpt-5.6-luna
Reasoning setting: high
Token limit including reasoning: 8,192 tokens
Output tokens: 3,164 tokens
Cost of the attempt: $0.0038
Response time: 49.9 s
API model reported in response: openai/gpt-5.6-luna
Pinned endpoint: OpenAI | openai/gpt-5.6-luna-20260709
Temperature: Provider default
Test protocol: community-visual-tests-v1
Technical checks: Passed, not a quality rating

Visually reviewed. This individual test does not establish overall model quality.

DeepSeek-V4.1-Flash

The response reached the token limit. No complete recording is available.

Oct 5, 2026 · 26 s · $0.0049

Not yet visually reviewed. A successful recording does not confirm correct physics or full compliance with the prompt.

Test details

Run date: Oct 5, 2026
API provider: openrouter
Requested API model: deepseek/deepseek-v4.1-flash
Reasoning setting: high
Token limit including reasoning: 8,192 tokens
Output tokens: 8,191 tokens
Cost of the attempt: $0.0049
Response time: 25.8 s
API model reported in response: deepseek/deepseek-v4.1-flash
Pinned endpoint: DeepSeek | deepseek/deepseek-v4.1-flash-20260910
Temperature: Provider default
Test protocol: community-visual-tests-v1
Technical checks: Failed

Earlier attempts (2)

Oct 2, 2026API request failedReasoning setting: high
Oct 1, 2026API request failedReasoning setting: high

SVG short test

One SVG task with the same fixed settings for every model, without hidden thinking.

Prompt

Return only one standalone SVG of a pelican riding a bicycle. Use viewBox="0 0 800 600", simple shapes, two clearly visible wheels and a recognizable pelican with a long beak. At most 30 SVG elements and 1800 characters. No Markdown, explanations, comments, scripts, animation, CSS or external resources.

GPT-5.6 Luna

Unavailable in this round

DeepSeek-V4.1-Flash

GPT-5.6 Luna and DeepSeek-V4.1-Flash in the overall index

Every model with a place in the overall index, both models of this matchup highlighted. You can add or hide other models.

Overall index (Gradually Meta-Score v4)

20 of 91 models

4 models in this view are measured at a lower reasoning effort than max in the overall index. Dashed column: lower reasoning effort than the top.

Claude Opus 5.5max126.1

Gemini 4 Carbonhigh125.7

Claude Fable 5.1max123.8

Claude Sonnet 5.5max122.8

Gemini 4 Argonhigh122.3

GPT-6 Astramax121.3

GPT-6.1 Solmax119.2

Muse Wonder Fullmax118.3

Claude Opus 5max117.2

Muse Wondermax113.4

Kimi K3.1 Promax112.9

Claude Fable 5max112.5

GPT-6 Solmax110.6

GPT-5.6 Solmax108.7

Muse Spark 1.3max107.4

Gemini 3.7 Flashhigh107.1

DeepSeek V4.1 Flash 1017max106.6

Gemini 3.8 Flashhigh106.5

DeepSeek-V4.1-Flashmax99.7

GPT-5.6 Lunamax93.9

Claude Opus 5.5

Gemini 4 Carbon

Claude Fable 5.1

Claude Sonnet 5.5

Gemini 4 Argon

GPT-6 Astra

GPT-6.1 Sol

Muse Wonder Full

Claude Opus 5

Muse Wonder

Kimi K3.1 Pro

Claude Fable 5

GPT-6 Sol

GPT-5.6 Sol

Muse Spark 1.3

Gemini 3.7 Flash

DeepSeek V4.1 Flash 1017

Gemini 3.8 Flash

DeepSeek-V4.1-Flash

GPT-5.6 Luna

Popularity compared

How often the models are searched for, used and mentioned, on a scale from 0 to 100.

GPT-5.6 Luna is more popular than DeepSeek-V4.1-Flash: 46 to 38 out of 100.

Over time

Weekly values from October 18, 2026 to November 5, 2026

GPT-5.6 LunaDeepSeek-V4.1-Flash

All data side by side

The better value is bold. For the overall index only when the gap is clear.

Metric	GPT-5.6 Luna	DeepSeek-V4.1-Flash
Intelligence
Overall indexpoints, 100 = reference group average, higher is better	93.9	99.7
Rank	Rank 38 of 91	Rank 29 of 91
Price per 1M tokens
Inputwhat you send	$0.20above 272,000 tokens: $0.40	$0.15
Outputwhat the model writes	$1.20above 272,000 tokens: $1.80	$0.60
Cost per task (own tests)average of 2 tests both passed	$0.01	$0.008
Speed
Tokens per second	94	176
Popularity
Popularitysearch, usage and mentions, 0 to 100	46Rank 161 of 875	38Rank 298 of 875
Technical
Context windowhow much text the model reads at once	1.05 M tokens≈ 1,580 A4 pages	1 M tokens≈ 1,500 A4 pages
Knowledge cutoff	Feb 16, 2026	n/a
Released	Jul 9, 2026	Sep 10, 2026
Availability	Active	Active
Open weights	no	yes

DeepSeek-V4.1-Flash: At peak times (Mon, Tue, Wed, Thu, Fri, 01:00-04:00 UTC; Mon, Tue, Wed, Thu, Fri, 06:00-10:00 UTC) the price is $0.30 input and $1.20 output per 1M tokens. The table shows the off-peak price.

Prices without caching or batch discounts · Page count as a rule of thumb from Artificial Analysis

Is there a better choice?

Successor of GPT-5.6 LunaGPT-6 Luna

Overall index 96.3, rank 35. $0.10 input and $0.50 output per 1M tokens.

DeepSeek-V4.1-Flash vs. GPT-6 Luna

Successor of DeepSeek-V4.1-FlashDeepSeek V4.1 Flash 1017

Overall index 106.6, rank 17. $0.15 input and $0.60 output per 1M tokens.

GPT-5.6 Luna vs. DeepSeek V4.1 Flash 1017

Cheaper than DeepSeek-V4.1-FlashClaude Haiku 5.5

Overall index 103.4, rank 20. $0.10 input and $0.50 output per 1M tokens.

DeepSeek-V4.1-Flash vs. Claude Haiku 5.5

Shared benchmarks

Single results from the same benchmark version, task and metric. A value is bold only with the same test setup and a clear gap.

A: GPT-5.6 LunaB: DeepSeek-V4.1-Flash

BioMysteryBenchOverall · higher is betterTest setup not documented as identical61.5%67.8%Vals AIAs of 11/05/2026

Code MigrationOverall · higher is betterTest setup not documented as identical44.6%45.6%Vals AIAs of 10/29/2026

CyberBench v1.1Overall · higher is betterTest setup not documented as identical73.6%73.7%Vals AIAs of 11/05/2026

EMBOverall · higher is betterTest setup not documented as identical67.1%57.2%Vals AIAs of 10/20/2026

Finance Agent (v2)Overall · higher is betterTest setup not documented as identical55%53.5%Vals AIAs of 10/29/2026

GDPval-AAOverall · higher is better1,468.24 elo1,600 eloArtificial AnalysisAs of 11/06/2026

IOIOverall · higher is betterTest setup not documented as identical61.8%40.3%Vals AIAs of 10/29/2026

ITBench-AA (SRE)Overall · higher is better40.3%46.9%Artificial AnalysisAs of 11/06/2026

Legal Research BenchOverall · All-pass · higher is betterTest setup not documented as identical36.5%41.3%Vals AIAs of 11/05/2026

LegalBenchOverall · higher is betterTest setup not documented as identical84%83.3%Vals AIAs of 10/13/2026

Show 54 more benchmarks

LMArena Text, Style ControlOverall · higher is betterTest setup not documented as identical1,4531,476LMArenaAs of 10/28/2026

LMArena Vision, Style ControlOverall · higher is betterTest setup not documented as identical1,2571,261LMArenaAs of 10/19/2026

MedCodeOverall · higher is betterTest setup not documented as identical42.4%41.2%Vals AIAs of 10/29/2026

MedScribeOverall · higher is betterTest setup not documented as identical84.4%85.5%Vals AIAs of 10/29/2026

MMMU-ProOverall · higher is better78.6%77.0%Artificial AnalysisAs of 11/06/2026

MysteryMechanismOverall · higher is betterTest setup not documented as identical14.4%21.2%Vals AIAs of 10/29/2026

ProgramBenchOverall · higher is betterTest setup not documented as identical0%0.5%Vals AIAs of 11/05/2026

ProofBench v1.1Overall · higher is betterTest setup not documented as identical60%54%Vals AIAs of 10/29/2026

Public Benefits Bench v1.1Overall · higher is betterTest setup not documented as identical61.2%64.3%Vals AIAs of 10/12/2026

SkillsBenchOverall · higher is betterTest setup not documented as identical60.4%69.8%Vals AIAs of 10/29/2026

SRE BenchOverall · higher is betterTest setup not documented as identical1.5%0.8%Vals AIAs of 11/05/2026

Tax Agent BenchOverall · higher is betterTest setup not documented as identical24.8%26.6%Vals AIAs of 11/05/2026

Terminal-Bench 4.0Overall · higher is betterTest setup not documented as identical11.6%19.7%Vals AIAs of 11/05/2026

Terminal-Bench-ScienceOverall · higher is betterTest setup not documented as identical0%0%Vals AIAs of 11/05/2026

Vals Science IndexOverall · higher is betterTest setup not documented as identical25.3%29.7%Vals AIAs of 10/25/2026

Vibe Code Bench 1-100Overall · higher is betterTest setup not documented as identical22.6%16.4%Vals AIAs of 10/29/2026

Vibe Code Bench v1.1Overall · higher is betterTest setup not documented as identical77.1%84.7%Vals AIAs of 11/05/2026

BioMysteryBenchHard · higher is betterTest setup not documented as identical27.5%21.6%Vals AIAs of 11/05/2026

Code MigrationCLI · higher is betterTest setup not documented as identical36.1%37.5%Vals AIAs of 10/29/2026

CyberBench v1.1Patch · higher is betterTest setup not documented as identical83.9%85.7%Vals AIAs of 11/05/2026

EMBComps · higher is betterTest setup not documented as identical54.6%43.9%Vals AIAs of 10/20/2026

Finance Agent (v2)Adjustments · higher is betterTest setup not documented as identical52%49.2%Vals AIAs of 10/29/2026

GDPval-AAOverall · higher is better927.84 elo1,348.71 eloArtificial AnalysisAs of 11/06/2026

IOIIOI 2024 · higher is betterTest setup not documented as identical85%42.7%Vals AIAs of 10/29/2026

Legal Research BenchAdministrative / Regulatory · All-pass · higher is betterTest setup not documented as identical48.5%53%Vals AIAs of 11/05/2026

LegalBenchConclusion Tasks · higher is betterTest setup not documented as identical88.9%89.1%Vals AIAs of 10/13/2026

LMArena Text, Style ControlBusiness, management and finance · higher is betterTest setup not documented as identical1,4621,478LMArenaAs of 10/28/2026

LMArena Vision, Style ControlChinese · higher is betterTest setup not documented as identical1,2781,311LMArenaAs of 10/19/2026

MMMU-ProOverall · higher is better60.3%65.6%Artificial AnalysisAs of 11/06/2026

ProgramBenchAlmost Resolved · higher is betterTest setup not documented as identical9%16.5%Vals AIAs of 11/05/2026

SkillsBenchNo Skills · higher is betterTest setup not documented as identical45.2%61.7%Vals AIAs of 10/29/2026

SRE BenchCapability Score · higher is betterTest setup not documented as identical6.4%14.2%Vals AIAs of 11/05/2026

Tax Agent BenchControversy & Precedence Analysis · higher is betterTest setup not documented as identical10%3.3%Vals AIAs of 11/05/2026

Terminal-Bench 4.0Hardware · higher is betterTest setup not documented as identical0%0%Vals AIAs of 11/05/2026

Terminal-Bench-ScienceEarth Sciences · higher is betterTest setup not documented as identical0%0%Vals AIAs of 11/05/2026

Vals Science IndexBioMysteryBench · higher is betterTest setup not documented as identical61.5%67.8%Vals AIAs of 10/25/2026

BioMysteryBenchHuman-solvable · higher is betterTest setup not documented as identical69.4%78.5%Vals AIAs of 11/05/2026

Code MigrationCOBOL · higher is betterTest setup not documented as identical70%70%Vals AIAs of 10/29/2026

CyberBench v1.1PoC · higher is betterTest setup not documented as identical63.3%61.7%Vals AIAs of 11/05/2026

EMBDataroom · higher is betterTest setup not documented as identical83%73.9%Vals AIAs of 10/20/2026

Finance Agent (v2)All-Pass · higher is betterTest setup not documented as identical43.3%41.1%Vals AIAs of 10/29/2026

IOIIOI 2025 · higher is betterTest setup not documented as identical49.8%35.2%Vals AIAs of 10/29/2026

Legal Research BenchAdministrative / Regulatory · Weighted partial credit · higher is betterTest setup not documented as identical87.5%89.1%Vals AIAs of 11/05/2026

LegalBenchInterpretation Tasks · higher is betterTest setup not documented as identical82.1%80.4%Vals AIAs of 10/13/2026

LMArena Text, Style ControlChinese · higher is betterTest setup not documented as identical1,4781,523LMArenaAs of 10/28/2026

LMArena Vision, Style ControlCreative writing with images · higher is betterTest setup not documented as identical1,2571,254LMArenaAs of 10/19/2026

ProgramBenchFully Resolved · higher is betterTest setup not documented as identical0%0.5%Vals AIAs of 11/05/2026

SkillsBenchWith Skills · higher is betterTest setup not documented as identical60.4%69.8%Vals AIAs of 10/29/2026

SRE BenchFully Solved · higher is betterTest setup not documented as identical1.5%0.8%Vals AIAs of 11/05/2026

Tax Agent BenchControversy & Precedence Analysis · higher is betterTest setup not documented as identical61.2%52.8%Vals AIAs of 11/05/2026

Terminal-Bench 4.0Media · higher is betterTest setup not documented as identical0%8.3%Vals AIAs of 11/05/2026

Terminal-Bench-ScienceEngineering Sciences · higher is betterTest setup not documented as identical0%0%Vals AIAs of 11/05/2026

Vals Science IndexMysteryMechanism · higher is betterTest setup not documented as identical14.4%21.2%Vals AIAs of 10/25/2026

Code MigrationCode Quality · higher is betterTest setup not documented as identical3.195 / 54.411 / 5Vals AIAs of 10/29/2026

More comparisons

Compare GPT-5.6 Luna and DeepSeek-V4.1-Flash with the models people compare them with most often.

GPT-5.6 Lunavs.DeepSeek-V4.1-Flash

Choose GPT-5.6 Luna for

Choose DeepSeek-V4.1-Flash for

Where is the difference?

Strengths at a glance

GPT-5.6 Luna vs. DeepSeek-V4.1-Flash at the same effort

Which model for which task?

See for yourself: same task, one attempt

Pelican on a bicycle

SVG short test

GPT-5.6 Luna

DeepSeek-V4.1-Flash

GPT-5.6 Luna and DeepSeek-V4.1-Flash in the overall index

Overall index (Gradually Meta-Score v4)

Popularity compared

Over time

All data side by side

Is there a better choice?

Shared benchmarks

More comparisons

Compare another matchup

Predecessor and successor

The most popular comparisons