Enter opens the highlighted result. Esc closes.

Claude Opus 5 vs DeepSeek V3.2

Claude Opus 5 leads DeepSeek V3.2 on coding, 83.0 to 30.0 on the Vibe Coding Index, and DeepSeek V3.2 costs 28.4x less per output token ($0.88 vs $25.00 per million). Updated October 21, 2026; 59 benchmarks have a result for both.

Replace Claude Opus 5 Replace DeepSeek V3.2
Claude Opus 5Anthropic
Coding
83.0
Overall
81.6
Output $/M
$25.00
59
DeepSeek V3.2DeepSeek
Coding
30.0
Overall
37.9
Output $/M
$0.88
0
vs boards won of 59 shared

Claude Opus 5 or DeepSeek V3.2: which to pick

Pick Claude Opus 5 for coding: it scores 83.0 on the Vibe Coding Index against 30.0 for DeepSeek V3.2.

Pick DeepSeek V3.2 for output tokens 28.4x cheaper ($0.88 vs $25.00 per million) and a lower cost per solved task ($0.99 against $1.33).

Model AClaude Opus 5Anthropic83.0/1006 independent benchmarks$1.33 per solved taskModel BDeepSeek V3.2DeepSeek30.0/1003 independent benchmarks$0.99 per solved task
Shared results
59
Claude Opus 5 only
72
DeepSeek V3.2 only
16
Like-for-like categories
8 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.

At a glance

Coding index
Claude Opus 5 83.0 from 6 benchmarks, DeepSeek V3.2 30.0 from 3 benchmarks.
Widest category
Agentic work, 54.0 points apart on average over 2 shared benchmarks.
Widest single gap
ARC-AGI-2: Claude Opus 5 90.4% against DeepSeek V3.2 4.0%, 86.4 points apart on a 0 to 100 scale.
Output price
Claude Opus 5 $25.00 against DeepSeek V3.2 $0.88 per million output tokens, 28.4x apart.
Context window
Claude Opus 5 takes 1M tokens, DeepSeek V3.2 164K.

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
CategoryClaude Opus 5BarsDeepSeek V3.2Ties
Coding90
Intelligence20
Agentic work20
Reasoning80
Math20
Knowledge80
Web design80
General200
All shared boards590
CodingLike-for-likeSix of these boards feed the Vibe Coding Index90
Coding benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
SWE-bench VerifiedNot measured70.0%Only DeepSeek V3.2 measured
SWE-bench MultilingualNot measured59.0%Only DeepSeek V3.2 measured
LiveBench coding90.8%Not measuredOnly Claude Opus 5 measured
LMArena WebDev16891362Claude Opus 5
AA Coding Index78.044.2Claude Opus 5
LiveCodeBench (Vals AI)89.0%80.7%Claude Opus 5
SWE-bench Verified (Vals AI)97.0%67.6%Claude Opus 5
Terminal-Bench 4 (Vals AI)53.5%Not measuredOnly Claude Opus 5 measured
Vibe Code Bench (Vals AI)88.4%5.1%Claude Opus 5
Code migration (Vals AI)57.5%Not measuredOnly Claude Opus 5 measured
ProgramBench (Vals AI)3.0%Not measuredOnly Claude Opus 5 measured
IOI (Vals AI)84.3%Not measuredOnly Claude Opus 5 measured
Terminal-Bench 2.1 (Vals AI)84.6%Not measuredOnly Claude Opus 5 measured
Vibe Code Bench 1-100 (Vals AI)28.5%Not measuredOnly Claude Opus 5 measured
SkillsBench (Vals AI)60.4%Not measuredOnly Claude Opus 5 measured
CyberBench (Vals AI)65.4%Not measuredOnly Claude Opus 5 measured
LiveCodeBench (Artificial Analysis run)Not measured86.2%Only DeepSeek V3.2 measured
SciCode56.4%Not measuredOnly Claude Opus 5 measured
Terminal-Bench Hard (Artificial Analysis run)Not measured35.6%Only DeepSeek V3.2 measured
Terminal-BenchNot measured39.6%Only DeepSeek V3.2 measured
FrontierCode53.4%Not measuredOnly Claude Opus 5 measured
FrontierSWE52.0%Not measuredOnly Claude Opus 5 measured
WeirdML91.8%Not measuredOnly Claude Opus 5 measured
ALE-Bench2164.6Not measuredOnly Claude Opus 5 measured
LMArena Text: coding15331453Claude Opus 5
LMArena Text: software and IT15271449Claude Opus 5
LMArena WebDev: React17111350Claude Opus 5
LMArena WebDev: HTML16801371Claude Opus 5
LMArena WebDev: from an image1662Not measuredOnly Claude Opus 5 measured
SciCode13.8%Not measuredOnly Claude Opus 5 measured
Neither model has a result on the other 9 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
AA Intelligence Index50.821.5Claude Opus 5
Epoch Capabilities Index163.9146.1Claude Opus 5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index20
Agentic work benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
AA Agentic Index56.5Not measuredOnly Claude Opus 5 measured
Design Arena: agents building web apps1266Not measuredOnly Claude Opus 5 measured
Design Arena: agents building full-stack apps1306Not measuredOnly Claude Opus 5 measured
Design Arena: agents building mobile apps1348Not measuredOnly Claude Opus 5 measured
Design Arena: agents building Android apps1254Not measuredOnly Claude Opus 5 measured
Design Arena: agents building Godot games1279Not measuredOnly Claude Opus 5 measured
Design Arena: agents building games1261Not measuredOnly Claude Opus 5 measured
Design Arena: PowerPoint slides1265Not measuredOnly Claude Opus 5 measured
LiveBench research88.9%Not measuredOnly Claude Opus 5 measured
LiveBench browser84.0%Not measuredOnly Claude Opus 5 measured
LiveBench long agentic62.1%Not measuredOnly Claude Opus 5 measured
SREBench (Vals AI)12.2%Not measuredOnly Claude Opus 5 measured
Finance Agent (Vals AI)58.6%Not measuredOnly Claude Opus 5 measured
Tax Agent Bench (Vals AI)45.5%Not measuredOnly Claude Opus 5 measured
Legal Research Bench (Vals AI)55.3%Not measuredOnly Claude Opus 5 measured
Excel Modeling Benchmark (Vals AI)73.6%Not measuredOnly Claude Opus 5 measured
BioMysteryBench (Vals AI)79.3%Not measuredOnly Claude Opus 5 measured
Terminal-Bench-Science (Vals AI)27.1%Not measuredOnly Claude Opus 5 measured
Vals Science Index47.9%Not measuredOnly Claude Opus 5 measured
Time Horizon Index (Vals AI)18.8%Not measuredOnly Claude Opus 5 measured
CUA-bench (Vals AI)9.0%Not measuredOnly Claude Opus 5 measured
τ²-Bench TelecomNot measured90.6%Only DeepSeek V3.2 measured
APEX-Agents65.8%21.3%Claude Opus 5
OSWorld 2.031.4%Not measuredOnly Claude Opus 5 measured
BALROG63.4%Not measuredOnly Claude Opus 5 measured
PostTrainBench35.0%Not measuredOnly Claude Opus 5 measured
Vending-Bench 2$11,182$1,034Claude Opus 5
LMArena Search1242Not measuredOnly Claude Opus 5 measured
Neither model has a result on the other 13 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index80
Reasoning benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
GPQA Diamond93.4%80.3%Claude Opus 5
LiveBench reasoning91.2%Not measuredOnly Claude Opus 5 measured
MysteryMechanism (Vals AI)37.4%Not measuredOnly Claude Opus 5 measured
MMMU Pro (Vals AI)89.9%Not measuredOnly Claude Opus 5 measured
GPQA Diamond (Artificial Analysis run)93.7%84.0%Claude Opus 5
Humanity's Last Exam (Artificial Analysis run)54.9%24.6%Claude Opus 5
GPQA Diamond (Epoch AI run)93.9%Not measuredOnly Claude Opus 5 measured
ARC-AGI97.5%57.0%Claude Opus 5
ARC-AGI-290.4%4.0%Claude Opus 5
SimpleBench80.6%Not measuredOnly Claude Opus 5 measured
Chess puzzles43.0%Not measuredOnly Claude Opus 5 measured
Mystery game puzzles59.0%Not measuredOnly Claude Opus 5 measured
EBR-bench45.7%Not measuredOnly Claude Opus 5 measured
LMArena Text: hard prompts15271434Claude Opus 5
LMArena Text: expert15541436Claude Opus 5
LMArena Text: hard prompts in English15231446Claude Opus 5
LMArena Vision: diagrams1348Not measuredOnly Claude Opus 5 measured
LMArena Vision: homework1338Not measuredOnly Claude Opus 5 measured
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index20
Math benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
AIMENot measured84.6%Only DeepSeek V3.2 measured
MGSM (Vals AI)Not measured86.0%Only DeepSeek V3.2 measured
AA Math IndexNot measured92.0Only DeepSeek V3.2 measured
AIME 2025 (Artificial Analysis run)Not measured92.0%Only DeepSeek V3.2 measured
OTIS Mock AIME98.9%Not measuredOnly Claude Opus 5 measured
FrontierMath, tiers 1 to 385.6%Not measuredOnly Claude Opus 5 measured
FrontierMath, tier 473.2%Not measuredOnly Claude Opus 5 measured
ProofBench99.0%Not measuredOnly Claude Opus 5 measured
LMArena Text: math15331435Claude Opus 5
LMArena Text: mathematical work15441440Claude Opus 5
Neither model has a result on the other 3 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index80
Knowledge benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
MMLU Pro91.6%84.9%Claude Opus 5
LegalBench (Vals AI)87.0%76.1%Claude Opus 5
CaseLaw (Vals AI)Not measured55.4%Only DeepSeek V3.2 measured
CorpFin (Vals AI)73.2%51.0%Claude Opus 5
TaxEval (Vals AI)75.1%68.2%Claude Opus 5
MortgageTax (Vals AI)72.1%Not measuredOnly Claude Opus 5 measured
MedQA (Vals AI)Not measured93.9%Only DeepSeek V3.2 measured
MedCode (Vals AI)63.6%Not measuredOnly Claude Opus 5 measured
MMLU Pro (Artificial Analysis run)Not measured86.2%Only DeepSeek V3.2 measured
SimpleQA Verified59.9%Not measuredOnly Claude Opus 5 measured
LMArena Text: science15311446Claude Opus 5
LMArena Text: medicine and healthcare15251440Claude Opus 5
LMArena Text: legal and government15301425Claude Opus 5
LMArena Text: business and finance15011410Claude Opus 5
Neither model has a result on the other 4 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index80
Web design benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
Design Arena: websites13091175Claude Opus 5
Design Arena: UI components13491149Claude Opus 5
Design Arena: game dev13381132Claude Opus 5
Design Arena: data viz13291158Claude Opus 5
Design Arena: 3D13351133Claude Opus 5
Design Arena: all code categories13241164Claude Opus 5
Design Arena: SVG13111032Claude Opus 5
Design Arena: ASCII art13521083Claude Opus 5
LiveBench design83.0%Not measuredOnly Claude Opus 5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index200
General benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
LiveBench data analysis74.6%Not measuredOnly Claude Opus 5 measured
Public Benefits Bench (Vals AI)76.9%Not measuredOnly Claude Opus 5 measured
SAGE (Vals AI)49.4%Not measuredOnly Claude Opus 5 measured
MedScribe (Vals AI)91.0%Not measuredOnly Claude Opus 5 measured
Vals Multimodal Index73.9%Not measuredOnly Claude Opus 5 measured
IFBenchNot measured60.7%Only DeepSeek V3.2 measured
Long context reasoning (AA-LCR)82.0%73.3%Claude Opus 5
DTBench97.9%85.6%Claude Opus 5
LMCA64.5%28.8%Claude Opus 5
CL-benchNot measured12.4%Only DeepSeek V3.2 measured
Furniture assembly60.8%Not measuredOnly Claude Opus 5 measured
CL-bench LifeNot measured7.4%Only DeepSeek V3.2 measured
LMArena Text15091425Claude Opus 5
LMArena Text: instruction following15181413Claude Opus 5
LMArena Text: creative writing14931400Claude Opus 5
LMArena Vision1320Not measuredOnly Claude Opus 5 measured
LMArena Text: English15081439Claude Opus 5
LMArena Text: non-English15021409Claude Opus 5
LMArena Text: Chinese15731455Claude Opus 5
LMArena Text: German15271413Claude Opus 5
LMArena Text: Japanese15201374Claude Opus 5
LMArena Text: Korean15171370Claude Opus 5
LMArena Text: French15191429Claude Opus 5
LMArena Text: Spanish15231424Claude Opus 5
LMArena Text: Russian15071424Claude Opus 5
LMArena Text: Polish15221419Claude Opus 5
LMArena Text: multi-turn15001427Claude Opus 5
LMArena Text: longer queries15171428Claude Opus 5
LMArena Text: writing and language14991407Claude Opus 5
LMArena Text: entertainment and media14811395Claude Opus 5
LMArena Vision: English1315Not measuredOnly Claude Opus 5 measured
LMArena Vision: Chinese1400Not measuredOnly Claude Opus 5 measured
LMArena Vision: reading text1329Not measuredOnly Claude Opus 5 measured
LMArena Vision: humor1293Not measuredOnly Claude Opus 5 measured
LMArena Vision: creative writing1313Not measuredOnly Claude Opus 5 measured
LMArena Document1516Not measuredOnly Claude Opus 5 measured
Neither model has a result on the other 10 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Opus 5 and DeepSeek V3.2
BenchmarkClaude Opus 5DeepSeek V3.2Leader
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

Claude Opus 5$0.018Fits in one request
DeepSeek V3.2$0.0007Fits in one request

DeepSeek V3.2 costs less

Repository review

50K input + 3K output tokens

Claude Opus 5$0.325Fits in one request
DeepSeek V3.2$0.017Fits in one request

DeepSeek V3.2 costs less

Agent loop

200K input + 10K output tokens

Claude Opus 5$1.25Fits in one request
DeepSeek V3.2$0.067Exceeds the context window

DeepSeek V3.2 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 Claude Opus 5 and DeepSeek V3.2
SpecificationAnthropic Claude Opus 5DeepSeek V3.2
MakerAnthropicDeepSeek
WeightsNo public weights listedOpen weights
ReleasedJuly 24, 2026December 1, 2025
Context window1M164K
Input $/Mtok$5.00$0.29
Output $/Mtok$25.00$0.88
Coding index83.030.0
Independent measurements behind it63
Cost per solved task$1.33$0.99
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Opus 5 better than DeepSeek V3.2 for coding?

Pick Claude Opus 5 for coding: it scores 83.0 on the Vibe Coding Index against 30.0 for DeepSeek V3.2. Pick DeepSeek V3.2 for output tokens 28.4x cheaper ($0.88 vs $25.00 per million) and a lower cost per solved task ($0.99 against $1.33).

Which is cheaper, Claude Opus 5 or DeepSeek V3.2?

Claude Opus 5 costs $5.00 per million input tokens and $25.00 per million output tokens; DeepSeek V3.2 costs $0.29 and $0.88. Per solved SWE-bench Verified task, Claude Opus 5 costs $1.33 and DeepSeek V3.2 $0.99.

Which has the bigger context window, Claude Opus 5 or DeepSeek V3.2?

Claude Opus 5, with 1M tokens against 164K.

How many benchmarks is the Claude Opus 5 vs DeepSeek V3.2 comparison based on?

59 benchmarks have a result for both models as of October 21, 2026. Claude Opus 5's coding index rests on 6 independent benchmarks and DeepSeek V3.2'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/claude-opus-5-vs-deepseek-v3-2" title="Claude Opus 5 vs DeepSeek V3.2, 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/claude-opus-5-vs-deepseek-v3-2">LLMBenchmark.ai, Claude Opus 5 vs DeepSeek V3.2</a></p>